enhancing cho cell productivity through the stable depletion of
TRANSCRIPT
Enhancing CHO cell productivity
through the stable depletion of
microRNA-23
A thesis submitted for the degree of Ph.D
Dublin City University
By
Paul S Kelly, B.Sc. (Hons)
The research work described in this thesis was performed under the
supervision of
Dr. Niall Barron and Prof. Martin Clynes
National Institute for Cellular Biotechnology
School of Biotechnology
Dublin City University
September 2014
i
I hereby certify that this material, which I now submit for assessment on the programme
of study leading to the award of Ph.D. is entirely my own work, and that I have exercised
reasonable care to ensure that the work is original, and does not to the best of my
knowledge breach any law of copyright, and has not been taken from the work of others
save and to the extent that such work has been cited and acknowledged within the text of
my work
Signed: _______________________ (Candidate) ID No.: 55036275
Date: _______________________
ii
This thesis is dedicated to my loving parents
Bernard & Paula
iii
Acknowledgements
First and foremost, I would like to thank my supervisor Dr. Niall Barron for his guidance
and patience over the last 5 years of my Ph.D. From the moment you said “Pull up a pew”,
I knew this was the project for me. Thank you for tolerating my constant science puns,
you know you love them. It’s a miR-acle you’ve put up with it for this long. For always
being available in times of crisis, the dreaded Guava, and for leaving the office door open
for any questions that I ever had. Most of all, thank you for reading every page that I put
on front of you and helping in the gruelling process of thesis edits, I promise the next one
will be shorter. For performance above and beyond the call of duty.
I would like to thank Professor Martin Clynes for pointing me in the direction to selecting
the right project for me. For your enthusiasm over the years, lord knows we all need it at
times, and the kind words of support. For your light-hearted emails, adding a little spice
to those dreaded PhD blues. You’ve done so much for me and so many others in my time
in the centre.
To my saviour, my god, the “bioinfor-magician”, Dr. Colin Clarke. Thank you for saving
my PhD when my laptop decided to throw in the towel near the end of writing up.
I would like to thank Dr. Clair Gallagher for her help at such a critical point in my project,
FACS sorting 2304 CHO clones, and for the long hours spent sitting at the Guava
compensating for GFP. Believe me, I am truly in your debt and a great deal of
“compensation” is coming your way.
I would like to thank Dr. Laura Breen for all of her help with the OROBOROS. My
apologies for unleashing the alternative name for the OROBOROS to the centre and your
responsibility of explaining its meaning. It will forever be remembered.
Thanks to Dr. Sinead Aherne for helping me out with my microarray work, even if the
results weren’t the Mae West.
To the proteomics crew, Dr. Paula Meleady, Michael Henry, Shane Kelly and Mark
Gallagher, a huge thanks for all the help with my label-free work and data analysis.
Navigation through the uncharted lands of proteomics was greatly appreciated.
I would also like to thank Dr. Jonathan Bones from NIBRT in UCD for all of the help
with glyco-profiling my antibodies and reformatting my graphs for corrections.
To my molecular biology partner in crime, BD-Griff, D.J. Griffindor or Alan “The Lad”
Griffith…….lad lad lad. To the chemistry group attempting to slowly infiltrate the
molecular biology lab; Zara Molphy, Creina Slator and Andreea Prisecaru, thanks for
mixing things up over the last couple of years. Thanks to the rest of the second floor office
crew both past and present; Andrew McCann, Deirdre O’Flynn, Damian Pollard, Martin
Power and my neighbour Gemma Moore (the spar trips will be missed).
iv
Carol McNamara, Gerladine Chawke, Mairead Callan and Yvonne for all of their support
over the years, for fending off the big bad wolves and for keeping us in the monies. I
would like to thank Gillian Smith and Joe Carrey for doing all of our prep room work.
I would like to thank everyone else along the way (this covers you) for being the ear or
shoulder from time to time and to those who asked the dreaded question “so, when are
you finishing up?”……**Shudder**.
Lastly, I would like to thank my family, especially my parents, Bernard and Paula Kelly
(no, I’m not named after my mother) for supporting me throughout my PhD and from
well before that. I know for a fact, life would have been so much more stressful without
you two behind me. I will never be able to repay your love and kindness but let this be
the start of it.
v
Table of contents
Abstract: Enhancing CHO cell productivity through the stable depletion of
microRNA-23…………………………………………………………………………...1
Abbreviations…………………………………………………………………………...2
Publications……………………………………………………………………………..6
Talks and Posters………………………………………………………………………. 7
1.0 Introduction…………………………………………………………………………8
1.1 Chinese hamster ovary (CHO) cells………………………………………………..9
1.1.1 Improving CHO cell productivity…………………………………………12
1.1.2 Media supplements and bioprocess regimes………………………………15
1.1.3 Waste metabolites………………………………………………………...16
1.1.4 The genomics era – Unravelling CHOs genetic knots…………………….17
1.1.5 Cellular engineering………………………………………………………19
1.1.5.1 Engineering CHO cell growth…………………………………..19
1.1.5.2 Anti-apoptosis engineering…………………………………….. 20
1.1.5.3 Anti-autophagy engineering…………………………………….22
1.1.5.4 Metabolic engineering…………………………………………..23
1.1.5.5 Secretion engineering…………………………………………...24
1.1.5.6 Chaperone engineering………………………………………….25
1.1.5.7 Unfolded-protein response engineering………………………...26
1.1.6 Has the maximum productive capacity of CHO reached its climax?...........27
1.2 microRNAs: Global regulators of cellular phenotype...........................................27
1.2.1 Discovery of microRNA………………………………………………….28
1.2.2 miRNA biogenesis and processing………………………………………..29
1.2.3 Genomic organisation of microRNA and nomenclature…………………..31
1.2.4 microRNA mode of action………………………………………………..34
1.2.5 Target prediction algorithms and online databases/tools………………….35
1.2.6 microRNAs in Chinese hamster ovary cells – A role in the bioprocess……37
1.2.6.1 miR-17~92 cluster………………………………………………41
1.2.6.2 miR-23~27~24 cluster…………………………………………..43
1.3 microRNAs: Tools for CHO cell engineering…………………………………….45
1.3.1 microRNA identification………………………………………………….45
1.3.2 Transient microRNA interference………………………………………...46
1.3.3 Stable microRNA interference……………………………………………48
1.3.4 microRNA sponge technology……………………………………………52
vi
1.4 Glyco-engineering…………………………………………………………………57
1.4.1 Antibody structure and effector function………………………………….59
1.4.2 Antibody glycoforms…………………………………………………….. 62
1.4.3 Glycan-manipulation…………………………………………………….. 65
Aims of thesis…………………………………………………………………………..68
2.0 Materials and Methods……………………………………………………………69
2.1 General Techniques of Cell Culture………………………………………………70
2.1.1 Ultrapure water……………………………………………………………70
2.1.2 Glassware…………………………………………………………………70
2.1.3 Cell culture cabinet………………………………………………………..70
2.1.4 Incubators…………………………………………………………………71
2.2 Subculture of cell lines…………………………………………………………….71
2.2.1 Anchorage-dependent cells (Monolayer)…………………………………72
2.2.2 Suspension cells…………………………………………………………..72
2.2.3 Cell counting and viability determination………………………………...72
2.2.3.1 Trypan blue exclusion method………………………………….72
2.2.3.2 Guava ViaCount® assay………………………………………...73
2.2.3.2.1 WorkEdit……………………………………………...74
2.2.3.2.2 EasyFit Analysis………………………………………75
2.2.3.3 Cedex XS Analyzer……………………………………………..76
2.2.4 Cryopreservation of cells………………………………………………….77
2.2.5 Thawing of cryopreserved cells…………………………………………...77
2.2.6 Apoptosis analysis………………………………………………………...78
2.3 Other Cell Culture Techniques…………………………………………………...78
2.3.1 Reconstitution/Rehydration of lyophilised miRNA………………………78
2.3.2 siRNA/miRNA transfection………………………………………………79
2.3.2.1 siSPORTTM NeoFXTM Transfecting Reagent…………………...79
2.3.2.2 INTERFERinTM transfection reagent…………………………...80
2.3.3 Transfection of plasmid DNA…………………………………………….80
2.3.3.1 Lipofectamine 2000……………………………………………. 80
2.3.3.2 TransIT®-2020 Transfection Reagent………………………….. 81
2.3.4 Limited-diltution/FACS-based cloning…………………………………...81
2.4 Molecular Biology Techniques……………………………………………………82
2.4.1 DNase treatment of RNA…………………………………………………82
2.4.2 High Capacity cDNA reverse transcription……………………………….82
2.4.3 Polymerase Chain Reaction (PCR)………………………………………..83
vii
2.4.4 Fast SYBR® Green Real-time qPCR……………………………………..85
2.4.5 Determination of DNA/RNA quantity and quality………………………..86
2.4.5.1 Bioanalyser/RNA……………………………………………….86
2.4.5.2 NanoDrop® Spectrophotometer………………………………... 87
2.5.4.3 Gel electrophoresis/RNA…………………………………….…88
2.4.6 Total RNA isolation using TRI ReagentTM………………………………..89
2.4.7 PCR based miRNA Taqman® Low Density Arrays (TLDA)……………...89
2.4.7.1 Reverse transcription……………………………………………91
2.4.7.2 Pre-amplification………………………………………………. 92
2.4.7.3 PCR……………………………………………………………..92
2.4.7.4 Data Analysis…………………………………………………...93
2.4.8 Singleplex qRT-PCR validation: Taqman® microRNA assays…………...94
2.4.8.1 Reverse transcription……………………………………………95
2.4.8.2 PCR……………………………………………………………..96
2.4.9 MessageAmpTM aRNA amplification kit…………………………………97
2.4.10 CHO-specific oligonucleotide arrays…………………………………..102
2.4.11 Gel and LB media/LB agar preparation………………………………...102
2.4.11.1 Agarose gel preparation…..…………………………………. 102
2.4.11.2 LB media preparation…..…………………………………….102
2.4.11.3 LB agar plate preparation…………………………………….103
2.4.12 Restriction Enzyme Digestion………………………………………… 103
2.4.13 Enzyme modification of linearized plasmid DNA……………..……….104
2.4.13.1 Alkaline Phosphatase (AP)………………………………..….104
2.4.13.2 Kinase treatment…………………………………………..….105
2.4.13.3 Klenow treatment…………………………………………….105
2.4.14 Gel extraction……………………………………………………..……107
2.4.15 PCR purification……………………………………………………..…108
2.4.16 Ligation………………………………………………………………...109
2.4.17 Transformation………..………………………………………………..111
2.4.18 DNA mini-prep of plasmid DNA………………..……………………..112
2.4.19 DNA midi/maxi-prep of plasmid DNA…..…………………………….113
2.4.20 Plasmid diagnostics…………………………………………………….114
2.5 Mitochondrial studies……………………………………………………….…...115
2.5.1 MitoTrackerTM Deep Red………………………………………………..115
2.5.2 OROBOROS Oxygraph-2k……………………………………………...115
2.5.3 Mitochondrial Isolation………………………………………………….116
2.6 Proteomic analysis………………………………………………………………..117
2.6.1 Bradford assay…………………………………………………………...117
2.6.2 Western Blot……………………………………………………………..117
2.6.3 2-D clean up kit………………………………………………………….118
2.6.4 Label-free sample digestion (Trypsin and Lys-c)…......…………………119
2.6.5 C-18 peptide purification………………………………………………...120
2.6.6 Label-free Mass Spectrometry…………………………………………..120
2.8.6.1 Identification of proteins from mass spectrometry data……….121
viii
2.8.6.2 Progenesis LCMS……………………………………………...121
2.6.7 SEAP assay……………………………………………………………... 123
2.6.8 ELISA………………………………………………………………….. 123
2.6.9 Glutamate assay…………………………………………………………125
2.6.10 Akta Prime IgG purification…………………………………………....125
2.6.11 Glycomic Characterisation of Antibodies……………………………...126
2.7 Biotinylated miRNA pull-down of mRNA………………………………………128
2.8 Statistical analysis……………………………………………………………….. 129
3.0 Results…………………………………………………………………………….130
3.1 Identification of microRNAs involved in the regulatory control of CHO cell
growth rate with subsequent phenotypic characterisation………………………...131
3.1.1 Identification of microRNAs involved in the regulation of CHO cell growth
rate and phenotypic characterisation…………………………..………………132
3.1.1.1 Differential miRNA expression profile of CHO cells with varying
growth rates using an integrated miRNA, mRNA and protein expression
analysis technique ………………………………………………….… 132
3.1.1.2 Cell culture and sampling……………………………………...133
3.1.1.3 microRNAs associated with growth……….…………………..133
3.1.2 Phenotypic validation through transient microRNA screening…………. 134
3.1.2.1 Cell culture and transfection conditions………………………. 134
3.1.2.2 Summary of transient miRNA screen………………………….135
3.2 The potential of miR-34a as an engineering target……………………………..138
3.2 Exploring the master “tumour suppressor” miR-34a as a potential CHO
engineering target…………………………………………………………………… 139
3.2.1 Transient knockdown of miR-34a potentially enhances CHO cell
growth…………………………………………………………………………139
3.2.2 Transient inhibition of miR-34a mildly inhibits the induction of
apoptosis………………………………………………………………………140
3.2.2.1 Apoptosis assay development………………………………… 141
3.2.2.2 Transient inhibition of miR-34a mildly inhibits apoptosis in CHO
cells induced to undergo cell death…………………………………….143
3.2.3 Stable depletion of miR-34 using microRNA sponge technology……….144
3.2.3.1 miR-34 sponge design and generation…………………………145
3.2.3.2 Confirmation of miR-34 sponge binding efficacy……………..150
3.2.3.3 miR-34a overexpression potentially inhibits conserved targets.152
3.2.3.4 miR-34 sponge influence on endogenous miR-34a and
CCND1………………………………………………………………..154
3.2.4 Evaluation of miR-34 depletion in CHO mixed pool populations………156
3.2.4.1 miR-34 depletion in CHO-K1-SEAP cells…………………….156
ix
3.2.4.2 miR-34 depletion in IgG-secreting CHO-1.14 cells…………...158
3.2.5 Generation of miR-34 sponge clones…………………………………….159
3.2.5.1 GFP fluorescence across clonal panels………………………...159
3.2.5.2 Impact of miR-34 depletion across a CHO-SEAP cell panel…..160
3.2.5.3 Impact of miR-34 depletion across a CHO-1.14 cell panel…….165
3.2.6 Phenotypic evaluation of miR-34 depleted clones in a 5 mL
volume…….…………………………………………………………………..170
3.2.6.1 Selected miR-34 depleted CHO-SEAP clones…………….......170
3.2.6.2 Selected miR-34 depleted CHO-1.14 clones…………………..171
3.2.7 miRNA-mediated interference of CHO cell antibody glycoforms……....175
3.3 Stable depletion of independent microRNA duplex components: miR-23b and
miR-23b*…………………………………………………………………………….. 180
3.3 CHO cell engineering using miR-23b and miR-23b*……………………………...181
3.3.1 Selection of miR-23b and miR-23b* for stable knockdown……………..181
3.3.2 miR-sponge vector design ans stable CHO cell line generation………….182
3.3.3 Investigation of GFP fluorescence………………………………………183
3.3.4 Analysis in mixed population……………………………………………185
3.3.4.1 Impact on bioprocess phenotypes in CHO-K1-SEAP: Batch and
Fed-batch……………………………………………………………...185
3.3.4.2 Impact of miR-23 and miR-23b* depletion on Mab producing
cells……………………………………………………………………189
3.3.5 Analysis of single cell clones in 24-well plate…………………………...191
3.3.5.1 GFP fluorescence across clonal panels………………………...193
3.3.5.2 Impact on bioprocess phenotypes across CHO-K1-SEAP clonal
panels………………………………………………………………….194
3.3.5.3 Impact on bioprocess phenotypes across CHO-1.14 clonal
panels………………………………………………………………….205
3.3.6 Investigation of miR-23 and miR-23b* depletion on selected clones in 5 mL
Batch, Fed-batch and Fed-batch Temperature shift……………………………215
3.3.6.1 miR-23 and miR-23b* sponge CHO-SEAP clones……………215
3.3.6.2 miR-23 and miR-23b* sponge CHO-1.14 clones……………...226
3.4 Investigation of the molecular impact of miR-23 depletion in CHO-SEAP
cells……………………………………………………………………………………230
3.4.1 miR-23 depletion improves CHO-SEAP specific productivity………….231
3.4.2 SEAP transcript levels appear elevated in miR-23 sponge clones……….231
3.4.3 SEAP mRNA stability is not a contributing factor to enhanced
productivity……………………………………………………………………233
3.4.4 Cell size is not a contributing factor to clonal differences in
productivity……………………………………………………………………235
3.4.5 miR-23 depleted CHO clones exhibited elevated glutamate levels……...236
3.4.6 miR-23 depletion impacts positively on CHO cell mitochondrial
function………………………………………………………………………..237
3.4.6.1 Mitochondrial function of intact rCHO cells…………………..239
x
3.4.6.2 Mitochondrial function of permeabilised rCHO cells………….243
3.4.7 Mitochondrial content…………………………………………………...247
3.4.8 Assessment of miR-23 target transcript levels…………………………...248
3.5 miR-23 target identification and pathway analysis……………………………..253
3.5 microRNA target identification……………………………………………………254
3.5.1 Identification of global mRNA targets through Biotinylated miRNA
tags…………………………………………………………………………….254
3.5.1.1 Evaluation of biotin-miRNA workflow………………………..256
3.5.1.2 Microarray identification of biotin-labelled miR-23b and miR-
23b* using biotin-labelled mimics…………………………………….260
3.5.2 Identification of global protein targets using Label-free mass
spectrometry………………………………………………………………….. 263
3.5.2.1 Differentially expressed proteins in miR-23 depleted clones… 263
3.5.2.2 High priority targets of miR-23……………………………….. 270
3.5.2.3 Gene Ontology analysis of differentially expressed proteins…..273
3.5.3 Identification of differential protein expression in isolated mitochondria.273
3.5.3.1 Differentially expressed proteins………………………………274
3.5.3.2 Gene Ontology analysis………………………………………..275
3.5.3.3 Potential targets of miR-23…………………………………….277
4.0 Discussion………………………………………………………………………... 282
4.1 CHO cell engineering using miR-23……………………………………………..283
4.1.1 Selection of miR-23 for stable depletion………………………………... 285
4.1.2 GFP as an effective reporter for clone selection and sponge
efficiency……………………………………………………………………...287
4.1.3 Stable depletion of miR-23 improves CHO-SEAP productivity…………288
4.1.4 miR-23 depletion improves CHO cell growth in a cell type-dependent
manner………………………………………………………………………...290
4.1.5 Clonal variation and culture conditions impact on CHO cell phenotypes..292
4.1.6 miR-23 depletion enhances CHO cell mitochondrial function at Complex I
and II…………………………………………………………………………..297
4.1.7 Summary of miR-23 depletion in CHO cells…………………………….303
4.2 Biotinylated miRNA mimics as a novel method for mRNA target
identification………………………………………………………………………….304
4.3 Analysis of the whole cell proteome reveals potential miR-23 regulated
process………………………………………………………………………………...306
4.3.1 LETM1…………………………………………………………………..306
4.3.2 14-3-3-eta (YWHAH)…………………………………………………...309
4.3.3 Calcium signalling……………………………………………………… 311
4.3.4 TCA cycle and metabolism……………………………………………... 313
xi
4.3.5 Combatting oxidative stress…………………………………………….. 319
4.4 Proteomic analysis of enriched mitochondria………………………………..…321
4.4.1 A deeper insight into mitochondrial function……………………………322
4.5 Potential bona fide miR-23 targets………………………………………………326
4.6 Integrated “omics” profile of CHO cells with varying growth rates…………...328
4.6.1 Identification of miRNAs associated with cellular growth………………329
4.6.2 Transient screening of miRNA……………………………………….… 331
4.6.2.1 miR-23b*……………………………………………………... 332
4.6.2.2 miR-34a/34a*………………………………………………….337
4.6.2.3 miR-181d/200a……………………………………………….. 340
4.6.2.4 miR-532-5p…………………………………………………....341
4.6.2.5 miR-639………………………………………………………. 342
4.6.2.6 miR-206………………………………………………………. 342
4.6.3 Final miRNA screen comments………………………………………….343
4.7 CHO cell engineering using miR-34……………………………………………..345
4.7.1 miR-34’s depletion in various CHO cell types…………………………..345
4.7.2 miRNA-based glycoform engineering of recombinant antibodies………349
4.8 CHO cell engineering using miR-23b*…………………………………………..352
4.9 miRNAs place in the bioprocess……………...…………………………………355
Conclusions…………………………………………………………………………...358
Future work…………………………………………………………………………..360
Bibliography………………………………………………………………………….363
Appendix……………………………………………………………………………...422
1
Abstract: Enhancing CHO cell productivity through the stable depletion of
microRNA-23
The Chinese hamster ovary cell (CHO) has been the “work horse” of the
biopharmaceutical industry for the past 2 decades. Their preeminent position is due
to their genetic tractability and capacity to produce high quality recombinant
proteins with human-like post-translational modifications. Productivity limitations,
however, have stimulated the application of various optimisation strategies to
further improve therapeutic product yields. Despite the marked improvements in
bioprocess design and media formulation, enhanced CHO cell performance is by no
means routine. microRNAs (miRNAs) offer an attractive alternative to genetic
engineering using a protein-coding gene due to their ability to regulate multiple
molecular networks simultaneously in addition to reducing the cells translational
burden.
Our lab identified various miRNAs whose expression was closely correlated with
CHO cell growth rate including the oncogenic miR-17~92 cluster and the ribosome-
enhancing miR-10a, as well as various growth-associated miRNAs such as miR-30e,
miR-206, miR-451 and miR-639. Transient characterisation of a set of these
candidates revealed several to negatively regulate growth and apoptosis such as
miR-34a/34a* and miR-23b*; positively influence growth including miR-15a-16-1
and miR-532-5p and potentially enhance specific productivity in the case of miR-
200a. CHO clones with enhanced growth attributes could achieve high cell densities
early in culture prior to the induction of a temperature-shift, a process control
strategy commonly used to enhance specific productivity but often at the cost of
growth.
A common trade-off for enhanced growth is often the reduction in CHO cell specific
productivity. Stable depletion of miR-23, implicated in B-cell differentiation, using
a miR-sponge decoy was observed to enhance recombinant protein yield by ~3-fold
without influencing CHO-SEAP cell growth in batch culture. Interestingly, these
high producing clones also demonstrated a ~30% increase in oxidative metabolism,
an energy state previously associated with productivity. The mitochondrial electron
transport system components, Complex I and II, were observed to be responsible
for this enhanced energy provision. Proteomic analysis identified potential targets
of miR-23 involved in the TCA cycle e.g. IDH1 and MDHA/B and related to
mitochondrial function e.g. LETM1.
Process adaptation to complement inherent clonal attributes can be exploited to
enhance production yields as evident in the case of a miR-23b*-depleted clone
achieving a 50% improved SEAP yield in a fed-batch culture with temperature shift
(31°C). Exploiting the role of miRNAs in the regulation of bioprocess-related
phenotypes (e.g. miR-34a and its role in antibody fucosylation) highlights the
versatility of these small molecules for industrial application, a prospect further
explored within this thesis.
2
Abbreviations
ADCC – Antibody-dependent cell cytotoxicity
AGO – Argonaut
AMPO – Aminopeptidase O
AON – Antisense oligonucleotide
ASO – Antisense Oligonucleotide
Asp – Asparagine
AT-III – Antithrombin III
ATCC – American Tissue Culture Collection
ATF4 – Activating transcription factor 4
CDK – Cyclin-dependent kinase
CDKI – Cyclin-dependent kinase inhibitors
ceRNA – Competitive endogenous RNA
CERT – Ceramide transfer protein
circRNA – Circular RNA
CHO – Chinese Hamster Ovary
CTGF – Connective tissue growth factor
DE – Differential Expression
DGCR8 – DiGeorge syndrome critical region gene 8
DNA – Deoxyribonucleic Acid
dsDNA – Double-stranded DNA
ELISA – Enzyme-Linked immunosorbent assay
EPO – Erythropoietin
ER – Endoplasmic reticulum
ERAD – ER-associated degradation
Exp5 – Exportin-5
FCCP – Trifluorocarbonylcyanide Phenylhydrazone
3
Fuc – Fucose
FUT8 – α-1,6 fucosyltransferase
GADD34 – Growth arrest and DNA damage inducible protein 34
Gal – Galactose
GlcNAc – N-acetylglucosamine
GLS – Glutaminase
GU – Glucose units
GnTIII – β1,4-N-acetylglucosaminyltransferase III
GSIS – Glucose-stimulated insulin secretion
GOI – Gene of Interest
GRN – Gene regulatory network
HCP – Host cell proteome
HIF-1 – Hypoxia-inducible factor 1
HPLC – High performance liquid chromatography
IFN – Interferon
IL – Interleukin
IRES – Internal Ribosome Entry Site
LC/MS – Liquid chromatography/Mass spectrometry
LDH – Lactate dehydrogenase
LDSCC – Limited-dilution single cell cloning
LNA – Locked nucleic acid
lncRNA – Long non-coding RNA
MBS – MicroRNA Binding Site
mRNA – Messenger RNA
miRNA – microRNA
miRNA* – microRNA* or microRNA (star) or microRNA passenger strand
MRE – MicroRNA Response Element
4
NC – Negative control
ncRNA – Non-coding RNA
NK – Natural killer
nt – Nucleotide
PAR-CLIP – Photoactivatable-Ribonucleoside-Enhanced Crosslinking and
Immunoprecipitation
PCD – Programmed cell death
PCR – Polymerase chain reaction
PDH – Pyruvate dehydrogenase
PDI – Protein disulfide isomerise
POX – Proline Oxidase
Pre-miRNA – Preliminary microRNA
Pri-miRNA – Primary microRNA
PRODH – Proline Dehydrogenase
PTGR – Post-transcriptional gene silencing
PVA – Polyvinyl Alcohol
qPCR – Quantitative Polymerase chain reaction
RBD – RNA binding domains
RISC – RNA-Induced Silencing Complex
RNA – Ribonucleic Acid
RNAi – RNA Interference
RNA pol – RNA polymerase
ROS – Reactive oxygen species
ROX – Residual Oxygen
RP – Ribosomal protein(Lempiainen, Shore 2009)
rpm – Revolutions per minute
RT – Reverse transcription
5
SEAP – Secreted Alkaline Phosphatase
shRNA – Short Hairpin RNA
SNARE – Soluble NSF attachment protein receptor
ssRNA – Single-stranded RNA
TAUT – Taurine transporter
TCA – Tricarboxylic acid cycle
TMB – 3,3’,5,5’ – Tetramethylbenzidine
TNFR – Tumour necrosis factor receptor
tPA – Tissue plasminogen activator
TPO – Thrombopoietin
TPS-1 – Thrombospondin-1
TGE – Transient Gene Expression
TRBP – TAR RNA binding protein
TU – Transcriptional unit
UPR – Unfolded protein response
UTR – Untranslated region
VCP – Vasolin-Containing Protein
XBP – X-box binding protein
XIAP – X-linked inhibitor of apoptosis
6
Publications
Kelly PS, Breen L, Mick H, Kelly S, Clynes M and Barron N 2014. Stable miR-23
depletion enhances recombinant protein yield from a Chinese hamster ovary cell line with
an impact on mitochondrial function. Manuscript in preparation.
Kelly PS, Clarke C, Clynes M and Barron N 2014. Bioprocess engineering: Micro-
managing CHO cell phenotypes. Pharmaceutical Bioprocessing – Future Science.
Accepted – August 2014 publication.
Sanchez N, Kelly P, Gallagher C, Lao N, Clarke C, Clynes M and Barron N 2013.
Sponge decoy vectors effectively deplete miR-7 activity to boost CHO cell longevity and
yield. Journal of Biotechnology. 2014 Mar 9(3):396-404.
Clarke C, Meleady P, Doolan P, Henry M, Kelly S, Aherne S, Sanchez N, Kelly P,
Kinsella P, Breen L, Madden S, Zhang L, Leonard M, Clynes M and Barron N 2012.
Investigation CHO cell growth rate via the application of omics level platforms conducted
simultaneously on a range of closely related sister clones with various growth rates. BMC
Genomics 2012 Nov 21;13(1):656.
Barron N, Sanchez N, Kelly P, Clynes N 2011. MicroRNAs: tiny targets for engineering
CHO cell phenotypes? Biotechnol Lett. 2011 Jan; 33(1):11-21.
7
Scientific talks and posters
IBC’s 10th Annual Cell Line Development & Engineering. September 8-10th, 2014.
Double Tree by Hilton Berkeley Marina, Berkeley CA, USA.
I was invited to present a talk entitled “Enhancing Chinese hamster ovary cell
productivity using microRNA-23” in Berkeley California on the current research
within our group.
European Society for Animal Cell Technology (ESACT) – 23rd biennial meeting
2013. Lille, France, 23rd-26th of June.
I presented a research poster focusing on an aspect of my work entitled
“MicroRNAs: Engineering tools to enhance CHO cell performance – The “Master
tumour suppressor” miR-34a” at this prestigious animal cell technology
conference with the underlying theme of “better cells for better health”.
European Society for Animal Cell Technology (ESACT) – UK conference, 9th-10th
of January 2013.
I presented a research based lecture entitled “MicroRNAs: Engineering tools to
enhance CHO cell performance” at the European Society for Animal Cell
Technology (ESACT) – UK conference, 9th-10th of January 2013 in the Imago
Conference Centre, Loughborough University, Leicestershire LE11 3TU, United
Kingdom.
Biotechnology in Action: Stem Cells & Tissue Engineering, Biopharmaceutical
Production and Cancer Biomarkers, 4th-6th of September 2012.
I presented a research poster entitled “MicroRNA based Chinese hamster ovary
cell engineering – miRNA 34a as a potential route towards improving industrial
cell culture?”
8
Chapter 1
Introduction
This introduction has been accepted for publication as a review focusing on the
current status of microRNAs in the field of Chinese hamster ovary cell engineering.
The review is entitled “Bioprocess engineering: Micro-managing the CHO cell
phenotype”, will be published in FUTURE SCIENCE – Pharmaceutical
Bioprocessing [August 2014] and was significantly contributed to by Dr. Niall Barron
and Dr. Colin Clarke.
9
1.1 Chinese hamster ovary (CHO) cells
The Chinese hamster ovary (CHO) cell is the primary mammalian cell line used in the
biopharmaceutical industry for the production of recombinant therapeutic proteins. They
are favoured over their microbial counterparts (bacteria, yeast etc.) for their ability to
produce high quality proteins with human like post-translational modifications,
glycosylation, ensuring high efficacy. Since the first FDA-approved recombinant protein
was manufactured in CHO, tissue plasminogen activator (tPA or Activase®) (Birch,
Arathoon 1990), over 70% of all recombinants are produced in this cell, earning them the
status of the “biological workhorse” of the biopharmaceutical industry (Jayapal et al.
2007). Of the 58 approved biopharmaceuticals (Table 1.1) between 2006 and 2010, 32
were manufactured in CHO with the 28 currently available antibodies constituting a large
fraction of the market (Ho, Tong & Yang 2013).
Optimisation of their performance has been an area of intense research over the last two
decades in order to enhance production yields and ultimately meet the global market
demand. Since the 1990’s, product titres have been improved ~20-fold primarily as a
result of expression vector improvements and the ability to isolate high producing clones
in addition to tailor-designed media formulations and the implementation of modified
bioprocess regimes. Such improvements have boosted the industry standard to 1-5 g/L
for monoclonal antibodies (Mabs) (Butler, Meneses-Acosta 2012, De Jesus, Wurm 2011).
However, such achievements are by no means routine and the recent emergence of
complex therapeutic molecules, such as fusion proteins and antibody fragments, present
new manufacturing challenges for the CHO cell platform.
Enhancing CHO cell performance has taken various forms including the implementation
of bioprocess regimes such as temperature shift (Butler 2005) and chemical
supplementation of sodium butyrate, discussed later (Hendrick et al. 2001b), to enhance
specific productivity in addition to genetic engineering strategies to compliment these
previous breakthroughs by overcoming side-effects associated with the bioprocess. Single
and multi-gene engineering processes have been exploited to target CHO cell phenotypes
critical to the bioprocess including cell cycle (Sunley, Butler 2010), apoptosis (Lee et al.
2013), metabolism (Jeon, Yu & Lee 2011) and secretion (Rahimpour et al. 2013).
10
Table 1.1: List of FDA/EU approved biologics produced in CHO
Selected list of FDA/EU approved biologics produced in CHO (1987-2012)
Product Type Therapeutic use Manufacturer Year of
approval
(FDA/EU)
Lucentis Anti-VEGF-A
mAb
Diabetic macular
edema
Genetech 2012
Yevroy
Anti-CTLA-4 mAb
Melanoma
Bristol-Myers-
Squibb
2011
Elonva Modified rh-FSH Controlled ovarian
stimulation
N.V. Organon 2010
Opgenra rh-BMP-7 Posterolateral
lumbar spinal
fusion
Howmedica
2009
Recothrom
rh-Thrombin
Control of minor
bleeding during
surgery
Zymogenetics
(Seattle)
2008
Retacrit
rh-EPO
Anemia associated
with chronic renal
failure
Hospira 2007
Vectibix Anti-EGFR mAb Metastatic
colorectal cancer
Amgen 2006
Naglazyme
N-
acetylgalactosamin
e-4-sulfatase
Mucopoly-
saccharidosis VI
Genzyme
2005
Luveris Luteinizing
hormone
Infertility Serono 2004
Xolair Anti-IgE mAb Moderate/Severe
Asthma
Genetech 2003
Humira Anti-TNFα mAb Rheumatoid
arthritis
Abbott 2002
Aranesp Erythropoietin
(engineered)
Anemia Amgen 2001
ReFacto Factor VIII Hemophilia A Wyeth 2000
Herceptin Anti-HER2 mAb Metastatic breast
cancer
Genetech 1998
Benefix Factor IX Hemophilia B Wyeth `1997
Avonex Interferon-β Relapsing multiple
sclerosis
Biogen Idec 1996
Cerezyme β-
glucocerebrosidase
Gaucher’s disease Genzyme 1994
Pulmozyme Deoxyribonuclease
I
Cystic fibrosis Genetech 1993
Epogen Erythropoietin Anemia Amgen 1989
Activase Tissue
plasminogen
activator
Acute myocardial
infarction
Genetech 1987
11
Engineering approaches have diversified to encompass the tailored manipulation of
antibody glycoforms for the enhancement of effector functions such as antibody-
dependent cell cytotoxicity (ADCC) (Jefferis 2009a), overviewed in greater detail in
section 1.4. By reducing core fucosylation of the biantennary-glycan on the Fc region of
an antibody, therapeutic potency can be heightened as described in the case of the major
blockbuster HerceptinTM (Junttila et al. 2010). Process manipulation such as those
described, have demonstrated improved volumetric or specific productivity (Qp) by
sustaining a prolonged production phase, manipulating bioenergetics pathways to reduce
toxic by-product accumulation, improving post-translational processing or indirectly
meeting global demand by decreasing patient dosages.
The emerging prominence of CHO has been marked by the publication of both the
Chinese hamster (Lewis et al. 2013) and CHO (Xu et al. 2011) genomes in addition to a
wealth of ‘omics data unravelling the molecular fingerprint of CHO cell genetics (Hackl
et al. 2011, Becker et al. 2011) and phenotypic characterisation (Clarke et al. 2012, Clarke
et al. 2011a). CHO-specific genomic support will focus cell line engineering, allowing
for the derivation away from the dependency on conservation with human and mouse
sequences. In addition, vast improvements in the development of a plethora of tools and
techniques for the manipulation of the expression networks in order to bend the CHO cell
to the specific requirements of product molecules. Sequence access has not only gained
insight into the variation of bioprocess characteristics in CHO such as post-translational
networks and apoptosis (Xu et al. 2011) but will help refine currently implemented vector
design such as recombinase-mediated cassette exchange (Hirano et al. 2011), zinc-finger
nucleases (Cost et al. 2010), TALENS (Joung, Sander 2013), endogenous CHO
promoters (Pontiller et al. 2008) and CRISPR-Cas (Mali, Esvelt & Church 2013).
With over 350 antibodies currently in the pipeline of pre-clinical trial (Reichert 2013,
Walsh 2010) development (Reichert 2013), the emphasis on the availability of robust
“off-the-shelf” engineering tools encompassing the genetic diversity of parental CHO cell
lines and product specificity remains a priority in animal cell technology. Reducing the
translational burden (Hackl, Borth & Grillari 2012) associated with single and multi-gene
engineering strategies is now being addressed via RNAi (Wu 2009b) with miRNAs
offering this advantage in addition to simultaneous pathway manipulation (Jadhav et al.
2013).
12
1.1.1 Improving CHO cell productivity
Process limitations associated with cell line stability and transcriptional efficiency of the
recombinant product has contributed to the low productivity profiles of recombinant CHO
(rCHO) cultures, a constraint significantly relieved by advancements in vector constructs
(Li et al. 2007). Random integration of the recombinant gene of interest (GOI) gives rise
to an unpredictable expression pattern due to positional effects (Zhou et al. 2010)
resulting in the extensive screening of large numbers of CHO clones without addressing
stability, an attribute influenced by the type of promoter used (Brooks et al. 2004). As
only 0.1% of the genome is considered “active” (Little 1993), the frequency of successful
transgene insertions into a transcriptional “hot-spot” is low, further abrogating the
isolation of high producing clones.
The majority of production platforms utilize mutant strains of CHO cells (Table 1.2) that
have both alleles of the dihydrofolate reductase (dhfr) gene either mutated or deleted.
DHFR catalyses the conversion of folic acid to tetrahydrofolate and is necessary for
biosynthetic pathways that produce glycine, purines and thymidylic acid (Cacciatore,
Chasin & Leonard 2010). To support growth, DHFR-deficient cells would require media
supplemented with glycine, hypoxyanthine and thymidine. Incorporation of selection
markers on transgenes has allowed for the isolation of high producing CHO clones
through removal of untransfected subpopulations in addition to successful transfectants
whose endogenous expression is low. Gene amplification in dhfr-CHO and gs-CHO
(glutamine synthase) has been applied in the industrial production of recombinant
pharmaceuticals (Cacciatore, Chasin & Leonard 2010). By including the DHFR gene on
a plasmid with the GOI, cells can be isolated with further stringent selection through the
addition of methotrexate (MTX), inhibitor of DHFR, selecting high producers or gene
amplified cells through gene duplication.
13
Table 1.2: Selected list of commercially available CHO cell lines
Cell Line Description
CHO-DHFR or CHO-
DUKX-B11
Dihydrofolate reductase negative clone derived from the
parental CHO cell line
CHO-K1 A subclone from the parental CHO cell line
CHO-MT+ A subclone from the parental CHO cell line
Pro-5 Proline auxotrophs and require medium containing 40 mg/L
of proline
CHO-SSF A subclone of CHO-DUKX-B11
CHO-protein free CHO cells adapted to grow in protein-free medium
SuperCHO CHO-K1 subclone expressing insulin, transferring and IGF-
1
ViggieCHO DUKX-B11 subclone adapted to grow in absence of
exogenous growth factors
Various selection markers are widely available (Table 1.3) in addition to the inclusion of
reporter genes such as green-fluorescent protein (GFP) used to isolate clones based on
fluorescent-activated cell sorting (FACS) achieving reported specific productivities of
38-fold increase when compared to normal selection procedures (Sleiman et al. 2008).
FACS has also been employed under various conditions such as MTX-tagging
(Yoshikawa et al. 2001) and secretion based on surface protein expression (DeMaria et
al. 2008).
Table 1.3: List of selection markers available in mammalian expression vectors
Selectable marker Selective reagent
Metabolic selectable marker
Dihydrofolate reductase (DHFR) Methionine sulphoximine (MSX)
Glutamine synthase (GS) Methotrexate (MTX)
Antibiotic selectable marker
Puromycin acetyltransferase Puromycin
Blasticidin deaminase Blastcidin
Histidinol dehydrogenase Histidinol
Hygromycin Phosphotransferase Hygromycin
Zeocin resistance gene Zeocin
Bleomycin resistance gene Bleomycin
Aminoglycoside Phosphotransferase Neomycin (G418) The above table was source from an excellent review on cell culture process development (Li et al.
2010).
Various recombinases such as Cre and Flp, which recognise loxP and FRT sequences,
respectively, are applied in various fields of (Hirano et al. 2011) biotechnology (Hirano
14
et al. 2011). The availability of the CHO genomic sequence draft has allowed the
identification and pre-tagging of transcriptional “hot-spots” using the FRT/loxP system
in combination with a weakened promoter (Zhou et al. 2010). This process was expanded
on to include multiple site-specific insertions (Kameyama et al. 2010) and the subsequent
successful generation of rCHO with a high copy number of antibody genes (Kawabe et
al. 2012). Alternative to targeting specific high transcriptionally active genomic locations,
transposable elements have been explored for their unbiased selection in genomic
insertion, ability to increase copy number and generate stable production cell lines (Ley
et al. 2013).
Promoter choice has been a double edge sword for the expression of heterologous genes
within mammalian cell factories ensuring high transcriptional activity on one side through
the use of viral promoters such as CMV and SV40 (Spenger et al. 2004) while
contributing to instability due to silencing via DNA methylation, a common host response
to foreign genetic material (Robertson, Jones 2000). The exploration of endogenous CHO
promoters has been an area of interest as a means of navigating around this host response
(Pontiller et al. 2008). CHEF-1 was identified to be constitutively expressed in several
CHO cell lines (Running Deer, Allison 2004), an ideal quality for driving transgene
expression such as house-keeping genes. However, native CHO promoters that are
responsive to certain conditions such as temperature shift offer an attractive alternative to
inducible systems such as the identified calcyclin promoter in CHO (Pontiller et al. 2008).
A vast array of vector modifications can facilitate the selection of high producing CHO
clones that circumvents the requirement of screening thousands of clones ensuring that
the clones ultimately being selected will demonstrate a stable, constitutive and high
producing phenotype once in the bioprocess. However, beyond this assurance, the
heterogeneity within CHO cell clones themselves (Porter et al. 2010) allow for various
performance phenotypes once introduced into the wide range production platforms in
addition to the multitude of insults associated with production runs such as hypoxia,
nutrient deprivation, shear stress and toxin build-up.
15
1.1.2 Media supplements and bioprocess regimes
Both media composition and bioprocess design have contributed the most over the past 2
decades to achieving gram per litre titres (De Jesus, Wurm 2011). Although capable of
attached culture, the adaptability to suspension culture has made CHO cells a desirable
mammalian cell type for the large-scale production of biopharmaceuticals. Fed-batch is
the most commonly employed bioprocess used today and is generally applied to processes
that supply material for phase 3 clinical trials and for the market (Wurm 2004). This type
of process enhances maximal cell densities, prolongs culture viability and improves
product yields, however, cell waste accumulation is the major limiting factor in this
process type. Perfusion (Ryll et al. 2000) and continuous culture (Bleckwenn, Shiloach
2004) culture types can be beneficial for producing labile recombinant products in
addition to potential toxic proteins as culture media is constantly removed and replenished
with cells remaining in the case of perfusion culture.
Low temperature cultivation of recombinant CHO cells has previously been associated
with sustained cell viability and increased productivity (Al-Fageeh et al. 2006) with the
fundamental addition of maintaining product quality (Yoon, Song & Lee 2003). The
reduction in growth rate associated with hypothermic growth has been linked to an
accumulation of cells in the G1 phase of the cell cycle (Hendrick et al. 2001a), a phase
previously characterised as highly transcriptionally active (Kumar, Gammell & Clynes
2007). Several studies have been undertaken to explore the molecular mechanisms that
are activated upon temperature shift in CHO cells that stimulate this prolonged production
phase and enhanced specific productivity including miRNA profiling (Gammell et al.
2007, Barron et al. 2011b) and temperature-sensitive regulatory genes (Kantardjieff et al.
2010).
Artificial induction of high CHO cell specific productivity can also be achieved with the
addition of chemical supplements such as sodium butyrate (NaBu) in addition to other
chemicals such DMSO (Fiore, Zanier & Degrassi 2002), pentanoic acid (Liu, Chu &
Hwang 2001) and growth factors (Zheng et al. 2006). NaBu has been observed to arrest
cellular proliferation through the accumulation of cells in the G1 phase (Hendrick et al.
2001b). Residence of cells in this stage of the cell cycle is associated with a high
transcriptional activity in addition to increased metabolism and cell size (Kumar,
Gammell & Clynes 2007). The mechanism of action of NaBu has been suggested towards
16
its ability to act as an inhibitor of histone deacetylation improving genomic accessibility
(Jiang, Sharfstein 2008) including its ability to suppress c-myc (Mariani et al. 2003) and
induce p21 (Kobayashi, Tan & Fleming 2004). One metabolic state previously affiliated
with an increase in specific productivity has been oxidative phosphorylation (Templeton
et al. 2013) and has been observed to be induced under exposure to NaBu treatment
(Ghorbaniaghdam, Henry & Jolicoeur 2013). A previous study has demonstrated NaBu
to stimulate the release of calcium from the endoplasmic reticulum (ER) (Sun et al.
2012a). The association of calcium signalling with energy metabolism through the TCA
cycle potentially highlights the driving force of NaBu’s influence on CHO cell
productivity.(Kaufman, Malhotra 2014) However, multiple side effects are associated
with the addition of NaBu including growth inhibition, induction or repression of gene
expression and the induction of apoptosis (Davie 2003).
1.1.3 Waste metabolites
All of the aforementioned process modifications including the environment of the
bioreactor itself impose a particular set of insults which makes the culture environment
undesirable such as oxygen deprivation (hypoxia), nutrient depletion, metabolic waste
product build-up (lactate and ammonia) and sheer stress. These stresses ultimately reduce
the accumulation of cell density, induce the onset of cell death ultimately reducing the
maximum potential yield in addition to interfering with product quality.
Dissolved oxygen (DO) has been demonstrated to effect cell metabolism and the
glycosylation patterns of the recombinant product in addition to causing cellular damage
through the production of reactive oxygen species (ROS) (Shigenaga, Hagen & Ames
1994). Culture pH has also been demonstrated to elicit adverse effects on cell growth,
metabolism, protein production and product quality (Yoon, Song & Lee 2003). This
increase in culture pH is due to lactate production (Tsao et al. 2005), however, the
addition of alkaline reagents such as Sodium Hydroxide (NaOH) to combat culture pH
can be more detrimental by increasing osmolality of the culture resulting in cell death
(Chotteau, Lindqvist 2012). Lactate is considered less of a concern compared to ammonia
build-up (Kim do et al. 2013) due to its reversed metabolic consumption late in culture
(Ma et al. 2009). The build-up of ammonia has been observed to affect the glycosylation
patterns of recombinant proteins thus influencing their quality and bioactivity (Donaldson
17
et al. 1999). Reports have focused on minimising the generation of these byproducts by
limiting the availability of the primary nutrient sources of mammalian energy
metabolism, glucose and glutamine. Practical systems approaches have included the
replacement of glutamine with glutamate (Hong, Cho & Yoon 2010) and glucose with
galactose (Altamirano et al. 2004) resulting in the reduced accumulation of toxic
byproducts and enhanced production of recombinant products with favourable
modifications.
The utility of temperature shift and chemical-based approaches is abrogated by the
unfavourable side-effects induced such as apoptosis and growth arrest which ultimately
reduce the density of producing cells. Although certain engineering options such as NaBu
do not offer a viable application to industrial processes due to its scalability, such
strategies do offer a means of identifying phenotype-related genetic switches potentially
associated with desirable phenotypes such as high specific productivity (Barron et al.
2011b, De Leon Gatti et al. 2007).
1.1.4 The genomics era - Unravelling CHOs genetic knots
Omics-based approaches, such as transcriptomics, proteomics and metabolomics, have
been used in the development of rCHO cell lines destined for the production process,
reviewed in (Kim, Kim & Lee 2012). Recent advancements in omics tools have allowed
for the detection of global changes in DNA/RNA, protein and metabolites giving a better
understanding of the mechanism associated with particular CHO phenotypes and/or
certain metabolites that correlate adversely/beneficially with CHO cell performance.
Omics-based analysis sets the basis for CHO cell engineering drawing on the single or
multiple genes derived from these profiles. Recent advances in the area of omics-based
bioengineering has presented the researcher with a wealth of insight into the mechanisms
that drive cellular processes, reviewed in (Datta, Linhardt & Sharfstein 2013).
Prior to the release of the CHO genomic sequence, non-CHO-derived DNA microarrays
for mouse and rat were successfully utilised (De Leon Gatti et al. 2007, Baik et al. 2006).
Comparative transcriptomic analysis using DNA micro-array technology has been carried
out in rCHO cells that exhibit enhanced cell-specific productivity (q) under induced
culture conditions such as low temperature culture (Baik et al. 2006, Yee, Gerdtzen & Hu
18
2009), high osmotic conditions (Shen et al. 2010) and Sodium Butyrate (NaBu) treatment
(De Leon Gatti et al. 2007). Similarly, an additional study sought to predict the cell-
specific production capacity of rCHO by identifying genes expressed across a panel of
CHO clones with varying levels of q (Clarke et al. 2011b) and growth rate (Clarke et al.
2011a, Doolan et al. 2013).
The global scale analysis of the post-transcriptional gene regulation (PTGR) mechanisms
inherent in a cell system has been made possible from genomic and proteomic profiling
(Kanitz, Gerber 2010) allowing for the identification of effector protein targets that
functionally contribute to a specific cellular phenotype. Several proteomic studies have
been carried within rCHO cell culture systems that have been induced either chemically
(NaBu) or environmentally (temperature shift) to exhibit enhanced cell specific
productivity (Kumar et al. 2008, Yee et al. 2008). In addition, shotgun proteomics was
utilised to identify biomarkers that could be used to predict the production capacity of
rCHO cells when comparing a high-producing cell line against a low producing (Nissom
et al. 2006, Carlage et al. 2009b). An array of proteomic-based profiles have been carried
out in order to characterise and identify potential engineering targets involved in
apoptosis induction (Wei et al. 2011), hyperosmotic pressure (Lee et al. 2003) and CHO
cells cultured in serum-free medium supplemented with hydrolysates (Kim et al. 2011).
Several other proteomic-based tools are available for the identification of large-scale
changes within the host cell proteome (HCP) increasing the plethora of potential targets
for CHO cell optimization. CHO-specific sequence databases have allowed for the high
confident identification of CHO-specific proteins (Meleady et al. 2012c) thus adding to
the growing wealth of experimentally validated resources.
The power of omics technology has accommodated the exploration of specific markers
associated with CHO cell phenotypes allowing for the potential of predicting CHO cell
performance within the bioprocess. Clonal variation is a well appreciated and problematic
attribute in cell line development resulting in performance differences of a single clone
across multiple bioprocesses (Porter et al. 2010). The ability of predicting a cells
performance based on an expression pattern of a set of prerequisite candidates will
mediate the selection of clones potentially better equipped with various attributes such as
specific productivity (Clarke et al. 2011b) or growth (Clarke et al. 2012). Extracellular
metabolite profiling will aid in media development and feeding strategies. Such studies
include the characterisation of changes in metabolite accumulation over the culture
19
process (Bradley et al. 2010)(Chong et al. 2009b), profiling of both intracellular and
extracellular metabolites in high producing versus low mAb-producing rCHO cells
(Chong et al. 2012) and metabolite accumulation correlated with increased caspase
activation and the onset of apoptosis (Chong et al. 2011).
The power of omics technology is evident from the above studies and have led to the
identification of priority targets whose genetic manipulation will further enhance the
CHO cell phenotype, a process that can be utilised to overcome the inherent adverse side-
effects of the bioprocess.
1.1.5 Cellular engineering
An additional layer of engineering exists for the improvement of rCHO cell
characteristics in relation to cell growth, recombinant protein production, improved
cellular longevity and optimal nutrient utilisation and metabolism. Both gene-based, RNA
interference (RNAi) based strategies, and a combination of both, have been successfully
implemented in the optimisation of CHO cell activity, reviewed in (Kim, Kim & Lee
2012) and (Wu 2009a). However, the reliance on a single component approach depends
strongly on the identification of critical targets that have the capacity to markedly
influence the cellular functions associated with recombinant protein expression in CHO
cells. Engineering of particular cellular processes covering multiple avenues are
described below.
1.1.5.1 Engineering CHO cell growth
A common feature of enhanced specific productivity (qP) conditions are perturbations
within the cell cycle leading to a G0/G1 phase arrest, resulting in enhanced protein
biosynthesis, mediated by physical (low temperature) and/or chemical (NaBu)
environmental stimuli (Sunley, Butler 2010). Based on this observation, anti-cell cycle
mediators such as p27KIP1 and p21CIP1 have been used in rCHO cell engineering to obtain
a similar effect without altering the culture environment. Fussenegger et al. achieved a 4-
fold increase in specific productivity in a secreted alkaline phosphatase (SEAP)
producing cell line through the transient expression of p27KIP1 and p21CIP1 under the
20
transcriptional control of a tetracycline-repressible system (Fussenegger, Mazur & Bailey
1997). Stable overexpression of these cyclin-dependent kinase inhibitors resulted in a
significant increase in specific productivity of 5-15 times for recombinant SEAP and
soluble intracellular adhesion molecule (sICAM) (Mazur et al. 1998, Meents et al. 2002).
A temperature shift-like phenotype was observed upon the transient overexpression of
the microRNA, miR-7, in CHO cells resulting in an increase in normalised productivity
(Barron et al. 2011b).
Although strategies such as those described above have led to the increase in normalised
CHO cell productivity, the inhibition of maximal CHO cell accumulation through the halt
of cell cycle progression results in no improvement to a decline in total volumetric titre.
Based on this consensus, the inverse relationship between growth rate and productivity
(Wang et al. 2006) must be coupled with engineering strategies that increase maximal
cell density before entering the “production phase”. Improvements in maximal cell
density and growth rate have been mediated by the overexpression of pro-cell cycle genes
such as the cyclin-dependent kinase, cyclin E, and the transcription factor E2F-1 (Jaluria
et al. 2007, Majors et al. 2008). Another study demonstrated a 70% increase in the
maximal cell density achieved through the overexpression of the c-myc transcription
factor which was observed to increase the cell cycle kinetics by ushering the cells into S-
phase at an accelerated rate (Kuystermans, Al-Rubeai 2009). Target identification
through omics-based approaches have been carried out to identify and validate several
gene targets involved in proliferation such as Vasolin-containing protein (VCP) (Doolan
et al. 2010) and requiem (Lim et al. 2010).
1.1.5.2 Anti-apoptosis engineering
As mentioned above, cell density is correlated with the maximum production capacity
during the bioprocess and with this in mind cell death is considered a critical issue to be
addressed as it effects the viable cell concentration thus impacting on product quality and
volumetric titre (Arden, Betenbaugh 2004). During rCHO cell culture, cell death
pathways can be triggered by a variety of stresses previously mentioned such as nutrient
depletion, hypoxia, accumulation of toxic metabolic byproducts, increased osmolality and
mechanical stress. The prevention of apoptosis has been recognised and established by
21
the generation of apoptosis-resistant rCHO cells as a beneficial engineering process to
enhance CHO cell performance.
It has already been well established that apoptosis in rCHO cells can be significantly
abrogated by the overexpression of anti-apoptotic proteins including Bcl-2, Bcl-xL and
Mcl-1 or the down-regulation of pro-apoptotic proteins such as Bak and Bax with the
observed increase in protein production (Cost et al. 2010, Majors et al. 2008, Majors et
al. 2009, Sung, Lee 2005, Chiang, Sisk 2005).
The apoptotic signalling cascade is mediated by a series of downstream proteolytic
enzymes composed of initiator and effector caspases that regulate the induction,
transduction and amplification of pro-apoptotic signals. Specific inhibition of the initiator
caspases, caspase-8/-9, through the stable expression of dominant negative (DN) mutants
(Yun et al. 2007) and inhibition of the effector caspases, caspase-3/-7, through plasmid
based expression of small interfering RNA (Kim, Lee 2002, Sung et al. 2007) has
demonstrated beneficial effects on cell growth and culture longevity. A further strategy
for suppressing caspase activity is through the overexpression of an intracellular caspase
inhibitor, X-linked inhibitor of apoptosis (XIAP) which has previously been demonstrated
in CHO to inhibit the activity of caspase-3, -7 and -9, offering protection against Sindbis
virus-induced apoptosis (Sauerwald, Betenbaugh & Oyler 2002a, Sauerwald, Oyler &
Betenbaugh 2003). Two independent studies of XIAP overexpression demonstrated the
contrary nature of its use in apoptosis inhibition and as a viable tool in CHO cell
engineering. The first exhibited no change in growth or productivity when CHO cells
expressing EPO were exposed to sodium butyrate despite the simultaneous reduced
activity of caspase-3, -7 and -9 (Kim, Kim & Lee 2009). The second conferred a
significant growth benefit and increased viable state in CHO cells that were serum
deprived through the reduced expression of caspase-3 (Liew et al. 2010). These mixed
results offer XIAP overexpression as a means of eliminating serum from the CHO cell
culture process, particularly during clonal development, while not as a combinatorial
strategy with chemical additives.
With the realisation that the apoptosis network is comprised of a complex network of
genes, several studies focused on the manipulation of simultaneous avenues in a multi-
gene approach. The expression of the mitochondrial pathway associated Bcl-xL and Aven
along with the additional expression of XIAP demonstrated additive effects in the
22
protection against cell death in mammalian cell culture (Sauerwald et al. 2006, Figueroa
et al. 2004). Apoptosis engineering has proved to be effective; however, there are
conflicting reports as to its influence on productivity. Some have reported negative and
positive impacts on CHO cell productivity. Nevertheless, combining such strategies with
chemical based methods, such as NaBu, have resulted in a significant increase in
recombinant therapeutic protein production (Sung et al. 2007) thus supporting the need
for the evaluation of the role of apoptosis pathway manipulation and target identification
in CHO.
1.1.5.3 Anti-autophagy engineering
Autophagy is a variant of PCD that takes place through a caspase-independent process in
response to both normal and stressful environmental conditions. Autophagy is a catabolic
process that acts in a homeostatic way by actively turning over cytoplasmic organelles
via lysosomal-mediated degradation. It has been observed that an increase in autophagy
induction can be triggered under stressful conditions such as nutrient deprivation and
accumulation of protein aggregates (Levine, Yuan 2005, Mizushima, Yoshimori &
Levine 2010). It is surmised that cell death may occur in the event of a prolonged
autophagic state as a result of excessive degradation of mitochondria or other vital cellular
constituents (Pattingre et al. 2005). Inhibition of apoptosis does not guarantee the
blocking of autophagy-mediated cell death. With the observation that autophagy was
induced late into the batch culture process of rCHO upon nutrient exhaustion (Hwang,
Lee 2008), the reverse was also noted with the reduction of apoptosis and autophagy upon
nutrient supplementation (Han et al. 2011). There are limitations in the online monitoring
of the induction of autophagy with the microtubule associated protein 1 light chain (LC3)
being the only bona fide autophagy marker (Kabeya et al. 2000). Autophagy progression
through FACS-based identification of this protein has been achieved (Eng et al. 2010)
with the turnover of long lived proteins offering a potential alternative (Klionsky, Cuervo
& Seglen 2007). CHO cell engineering to over-express Bcl-xL or the active form of Akt
were observed to delay the onset of apoptosis and autophagy as a result of nutrient
depletion (Kim et al. 2009a, Hwang, Lee 2009).
23
1.1.5.4 Metabolic engineering
Normal cell growth and metabolic turnover of primary energy sources such as glucose
and glutamine can result in the accumulation of byproducts including lactate and
ammonia which have been characterised as toxic and impede cellular proliferation and
product quantity and quality. Although feeding strategies and the availability of primary
nutrient sources have been demonstrated to reduce the production and build-up of such
byproducts in CHO cell culture, interference on a genetic level has proven to be a viable
option.
Continuously cultured rCHO cell lines have the metabolic disadvantage of completely
oxidising glucose to CO2 and H2O (Neermann, Wagner 1996) resulting in conversion of
glucose to pyruvate and ultimately lactate, the “Warburg effect”. Increased acidification
due to elevated lactate levels is mediated by the interconversion of pyruvate to lactate by
the catalytic enzyme lactate dehydrogenase (LDH). Genetic manipulation of LDH
activity has already been demonstrated to reduce the specific production of lactate in both
hybridoma and CHO cell lines (Chen et al. 2001, Jeong et al. 2001). siRNA-mediated
suppression of LDH-A, the subunit predominantly associated with anaerobic metabolism
and pyruvate reduction (Baumgart et al. 1996), yielded a 54-87% reduction in specific
glucose consumption along with a 45-79% reduction in lactate production with no
impairment in cellular proliferation and hTPO production (Kim, Lee 2007). An additional
layer of glucose metabolism is the conversion of pyruvate to acetyl-CoA for further
metabolism in the tricarboxylic acid (TCA) cycle. The enzyme, pyruvate dehydrogenase
(PDH), is responsible for this interconversion and its activity is regulated by PDH-Kinase
(K) and PDH-Phosphatase (P). It is for this reason that pyruvate is considered the branch
point that determines whether the consumed carbon source is excreted as lactate or
metabolised further in the TCA cycle. Zhou and colleagues explored the use of a dual
siRNA-expression vector against LDH-A and PDHK as a means of further reducing
lactate production while also maximising energy generation from the TCA cycle. CHO
cells exhibiting limited activity of LDH-A and increased activity of PDH showed reduced
lactate production (90%) with the addition of both increased cell specific productivity
(75%) and volumetric productivity (68%) (Zhou et al. 2011b). Specific glucose and
lactate production has been directly correlated with mAb production allowing the
potential to assess and predict CHO cell productivity based on metabolic turnover.
Additionally, a high level of mitochondrial dehydrogenase activity was observed
24
confirming that more efficient utilization of carbon sources through energy release from
the TCA cycle impacts positively on the biosynthesis capacity of CHO (Chen et al. 2012).
Another method explored was the controlled utilization of an alternative energy source to
glucose, fructose. Overexpression of the fructose-specific transporter (GLUT5) allowed
for the moderate uptake of fructose in high fructose containing environments, avoiding
the overflow of excessive carbon and lactate. The same clones cultured in glucose
environments demonstrated high sugar consumption and lactate production. The reduced
production of lactate in these GLUT5 clones resulted in an increased final cell
concentration but was only observed for clones that exhibited an appropriate expression
of this fructose transporter (Wlaschin, Hu 2007). Prolonged culture viability and
enhanced productivity was observed in the instance of stable overexpression of the
taurine-transporter (TAUT). Enhanced consumption of particular amino acids was
evident, taurine and β-alanine, which are known to function as organic osmolytes thus
inducing apoptosis upon their accumulation (Tabuchi et al. 2010).
To understand the bioprocess more intimately, a metabolomics-based approach was
carried out to identify the accumulation or depletion of particular metabolites at specific
time points within the culture using HPLC and LC/MS (Chong et al. 2009a). These
findings gave rise to the potential of using such a strategy for guiding media design and
fed-batch processes. A similar study, focused on the identification of novel metabolites,
oxidized glutathione, that potentially induced apoptosis through correlation with elevated
caspase activity (Chong et al. 2011). Metabolite characterisation studies such as these will
accommodate the controlled manipulation of CHO cell metabolism thus allowing
maximum performance.
1.1.5.5 Secretion engineering
Soluble NSF (N-ethylmaleimide-sensitive factor) attachment receptor, SNARE, proteins
are anchored to both transport vesicles and their target membrane, trigger membrane
fusion and are crucial in the exocytic pathway (Jahn, Scheller 2006). Vesicle trafficking
and SNARE-mediated membrane fusion is regulated by the interaction with Sec/Munc18
(SM). Several efforts have enhanced the production capacity of CHO cells by overcoming
this bottleneck associated with a saturated secretory pathway. Overexpression of
25
components related to the secretory network such as SNAREs (SNAP-23 and VAMP8)
and the SM proteins (Sly1 and Munc18c) have been exercised with a positive impact on
the secretory capacity of rCHO (Peng, Abellan & Fussenegger 2011, Peng, Fussenegger
2009). An alternative strategy implemented to enhance the secretory capacity of tPA-
producing rCHO cells was the generation of a mutant CHO cell line that possessed a
phosphorylation-resistant copy of ceramide-transfer protein (CERT), which efficiently
transfers ceramide from the endoplasmic reticulum (ER) to the Golgi. Specific
productivity was observed to be elevated by 35% (Rahimpour et al. 2013).
1.1.5.6 Chaperone engineering
A study identifying the irrelevance of secreted protein levels versus increasing gene copy
number/strong promoter led to the understanding that the bottleneck in enhancing specific
productivity lies at the translational and/or post-translational level (Mohan et al. 2008a).
To relieve the cells of this blockage, engineering of the chaperone machinery networks
have been undertaken. Protein disulfide isomerise (PDI) is predominantly located in the
ER and has been found to catalyse disulfide bond formation, assisting in the folding of
newly synthesised proteins (Freedman, Hirst & Tuite 1994, Gething, Sambrook 1992)
and acting as a molecular chaperone associated with miss-folded proteins in the ER (Cai,
Wang & Tsou 1994, Quan, Fan & Wang 1995). From an engineering perspective,
increased PDI activity in bacterial, yeast and insect hosts has been reported to increase
the secretion and activity of heterologous proteins containing disulfide bridges (Davis et
al. 2000). Overexpression of PDI in rCHO was observed to positively influence the
production capacity of antibody-producing CHO cells, but not in the case of TPO, IL-15
and tumor necrosis factor receptor:Fc fusion protein (Davis et al. 2000, Borth et al. 2005,
Mohan et al. 2007). Although antibody productivity was observed to be increased,
intracellular retention and co localisation in the ER was identified to be the contributing
factor to the reduction in TNFR:Fc. In a separate study, an isoform of PDI, ERp57, was
over-expressed to increase the levels of TPO secretion by 2.1 fold without any aberrant
effects on CHO cell growth (Hwang, Chung & Lee 2003). These reports give a mixed
view on the benefits of PDI overexpression as an engineering target with cell line,
expression system and target therapeutic protein playing a significant part. Simultaneous
overexpression of the chaperones calnexin and calreticulin exhibited a 1.9 fold increase
26
in specific TPO production with no change in it in-vivo biological activity (Chung et al.
2004). One proposed engineering strategy when exploiting the cells own protein
chaperone network would be to co-express several chaperones concomitantly in a
functionally meaningful ratio (Mohan et al. 2008b).
1.1.5.7 Unfolded-protein response engineering
With the identification of ER-associated chaperones, endoplasmin and PDI, being highly
correlated with CHO cell productivity (Smales et al. 2004) in addition to shotgun
proteomics revealing a positive correlation in productivity with the expression of the
molecular chaperone BiP (Carlage et al. 2009a), it was evident that the ER played a major
role in protein processing and the transmission of signals. However, accumulation of
unfolded proteins in the ER lumen induces a response named unfolded-protein response
(UPR). This response ultimately leads to the transcriptional activation of numerous
factors involved in the folding process, such as molecular chaperones, foldases (Travers
et al. 2000) and also genes involved in the ER-associated degradation (ERAD) pathway
(Casagrande et al. 2000). Prolonged or excessive UPR activation has been observed to
induce pancreatic beta cell death (Eizirik, Miani & Cardozo 2013) thus possessing the
potential to interfere with the viable nature of CHO along with offering an engineering
route to overcome this secretory bottleneck.
Reprogramming the post-translational processing machinery to overcome production
bottlenecks was demonstrated by the overexpression, both transient and stable, of X-box
binding protein 1 in transgenic CHO-K1 (Tigges, Fussenegger 2006). A spliced variant
(XBP-1s) was revealed to increase the specific production capacity of various therapeutic
proteins over its non-beneficial unspliced counter-part XBP-1u. Additionally, a further
study confirmed the utility of XBP-1s overexpression in CHO whereby a secretory
bottleneck had been reached with a 2.5 fold increase in specific productivity being
demonstrated for recombinant proteins such as mAb, erythropoietin (EPO) and interferon
γ (IFNγ) (Ku et al. 2008). Suspension CHO cell cultures maintained in a fed-batch format
exhibited a 40% increase in antibody titre with appropriate bioactivity and post-
translational modifications under stable XBP-1 expression (Becker et al. 2008).
27
In instances where the overexpression of XBP-1 failed to enhance the production of
recombinant proteins such as antithrombin III (AT-III), the constitutive activation of the
transcription factors ATF4 (Activating transcription factor 4) and GADD34 (Growth
arrest and DNA damage inducible protein 34) demonstrated improved titre of
recombinant AT-III (Ohya et al. 2008a, Omasa et al. 2008).
1.1.6 Has the maximum productive capacity of CHO reached its climax?
With such marked advancements in the area of recombinant CHO cell technology being
seen over the past 25 years, will further improvements in specific and volumetric
productivity be seen in the near future? Over the past 25-30 years, volumetric yields for
recombinant protein production have increased by about 20-fold primarily due to the
improvements in media and bioprocess design. It has been suggested (De Jesus, Wurm
2011) that a biological limit has been achieved at about 100 pg/cell/day in the area of
specific productivity. However, this is by no means routine and volumetric yields could
still be pushed higher, possibly 2-5-fold, by further increasing the maximal cell density
achieved within the bioreactor (De Jesus, Wurm 2011). By combining successful
bioprocess feeding strategies with that of genetically modified rCHO hosts, breaking the
current limitations of the bioprocess is a tangible goal. Focusing on the lesser explored
but emerging field of miRNAs, will not only minimise the translational burden associated
with gene-based genetic engineering strategies (Hackl, Borth & Grillari 2012) but will
fine tune global gene expression edging the CHO cell towards a more favourable
phenotype (Clarke et al. 2012). microRNAs (miRNAs) have been suggested as alternative
engineering tools in the past (Muller, Katinger & Grillari 2008, Wu 2009, Barron et al.
2011e) and in a short time have come a long way, reviewed in (Jadhav et al. 2013).
1.2 microRNAs: Global regulators of cellular phenotypes
MicroRNAs (miRNAs) are endogenous, small (~22 nt), single-stranded and non-coding
RNA (ncRNA) molecules that play an important functional role in the global regulation
of gene expression (Bartel 2004). Gene expression is exerted at the post-translational level
whereby mature miRNAs target mRNAs for translational inhibition or mRNA cleavage
depending on the level of complementarity between the miRNA and the 3’-untranslated
28
region (UTR). In some instances rapid mRNA decay has been attributed to accelerated
deadenylation via miRNA (Wu, Fan & Belasco 2006) with isolated occurrences of
enhanced translation of ribosomal proteins through interaction with their 5’UTRs (Orom,
Nielsen & Lund 2008).
It is predicted that miRNAs have the potential to target up to ~30% of protein coding
genes (Filipowicz, Bhattacharyya & Sonenberg 2008). The ability of a single miRNA to
potentially target hundreds of mRNA transcripts contributes to their complex nature (Chi,
Hannon & Darnell 2012). Conservation of the “seed” (nucleotides 2-8) sequence dictates
target specificity. Multiple miRNA-Recognition Elements (MREs) of a single miRNA
can be found within the 3’ UTR of several mRNA transcripts or several MREs belonging
to different miRNAs can lie within the same UTR giving rise to synergistic miRNAs
(Shukla, Singh & Barik 2011).
As of 2014, over 28,645 miRNAs have been discovered and logged in the miRBase
database with 1,881 of these being included in the human genome. miRNAs are involved
in a wide range of molecular pathways such as cellular proliferation (Dong et al. 2012),
apoptosis (Xie et al. 2012), fat metabolism (Flowers, Froelicher & Aouizerat 2012),
glucose-induced insulin secretion (Xia et al. 2011) and developmental timing (Carrington,
Ambros 2003). The conservation of 3’UTR miRNA seed sites across species (Friedman
et al. 2009) implicate these small RNA based gene regulators as prime targets for CHO
cell engineering and therapeutic intervention.
1.2.1 Discovery of microRNAs
The very first miRNA, lin-4, was identified in the nematode Caenorhabditis elegans in
1993 whilst studying its post-embryonic developmental timing. It was observed that the
transcription of lin-4 negatively regulated the levels of the LIN-14 protein at the L1 stage
of larval development. Ultimately, non-coding lin-4 transcript encoded for two small
mRNAs of 22 and 61 nucleotides that possessed complementary sequences to those found
in the 3’UTR of the lin-14 mRNA (Lee, Feinbaum & Ambros 1993). However, the lack
of conservation of lin-4 and lin-14 with any species other than nematodes suggested that
this was a “nematode specific curiosity” (Ambros 2008, Lee, Feinbaum & Ambros 2004,
Ruvkun, Giusto 1989).
29
A second and historically significant heterochronic gene switch was identified in the same
species, let-7, which was shown to code for a 21 nt RNA that possessed complementary
elements in the 3’UTR of its gene target lin-41 (Reinhart et al. 2000). Binding similarities
of let-7 to lin-4 was confirmed with imperfect complementary sequence binding located
within the 3’UTR of the lin-41 mRNA (Vella et al. 2004, Bagga et al. 2005). The
conservation of let-7 across several species such as vertebrate, ascidian, hemichordate,
mollusc, annelid and arthropod lead to the characterisation and identification of miRNAs
as a unique family of short interfering RNAs involved in species development,
homeostasis and disease (Pasquinelli et al. 2000). The multiple mRNA targets of the lin-
4 miRNA (lin-14 and lin-28) coupled with the predicted crossover target of let-7 with lin-
28 was the first indication that miRNAs could act both synergistically and independently
in numerous molecular networks (Reinhart et al. 2000, Moss, Lee & Ambros 1997).
1.2.2 miRNA Biogenesis and Processing
miRNAs are transcribed by RNA polymerase II, or RNA pol III, (Fig. 1.1 1) into long
primary (pri-) transcripts of several thousand bases, kilo-bases (kbs) (Lee, Feinbaum &
Ambros 2004, Steel, Sanghvi 2012). The removal of a shorter (~70 nt) hairpin stem loop
structure termed the preliminary miRNA (pre-miRNA) is performed by the
Microprocessing complex (Drosha-DGCR8, Fig. 1.1 2). Pri-miRNA processing and
recognition is mediated by the RNase type III enzyme Drosha and its double-stranded
RNA-binding adaptor protein DGCR8 (DiGeorge Syndrome Critical Region 8). Drosha
forms an intramolecular dimer whereby its two RNase III domains cut the 3’ and 5’ strand
independently of each other yielding a RNase III characteristic 2 nt-3’ overhang (Han et
al. 2004). Not only do the two double-stranded RNA binding domains of DGCR8
recognize the single-stranded RNA extensions flanking the base of the stem thereby
directing Drosha cleavage ~11 bp into the duplex from the ss-RNA/ds-RNA junction
(Zeng, Cullen 2005, Han et al. 2006) but the protein-protein interactions through
DGCR8/Drosha in the Microprocessor complex serve to stabilize the Drosha enzyme
(Han et al. 2006, Yeom et al. 2006).
Nucleo-cytoplasmic export of these newly processed pre-miRNAs is mediated by the
nuclear receptor Exportin-5 (Exp5) in a Ran-GTP dependent manner (Wang et al. 2011b),
(Fig. 1.1 3). Structural recognition of the pre-miRNAs stem duplex is weakly bound by
30
Exportin-5 with the identification of the 2 nt 3’ overhang being the pre-requisite for the
pre-miRNA recognition by Exp5 (Bohnsack, Czaplinski & Gorlich 2004, Lee et al. 2011).
In addition to the RNase III-specific 3’ overhang being the determining factor for miRNA
nuclear export, the pre-miRNA/Expotin-5/Ran-GTP complex also inhibits the
exonucleolytic degradation of the bound pre-miRNA (Zeng, Cullen 2004) thus stabilizing
and shuttling efficiently processed and folded miRNAs for further processing.
Figure 1.1: miRNA Biogenesis and Processing
Figure 1.1: 1) miRNA biogenesis begins with the transcriptional initiation by RNA pol II/III yielding
a primary miRNA transcript (pri-miRNA). 2) The Microprocessor complex containing
Drosha/DGCR8 cleaves the overhanging single stranded RNA liberating a ~70 nt Pre-miRNA
hairpin. 3) Nuclear removal of the pre-miRNA is mediated by Exportin-5 in a Ran-GTP dependent
fashion. 4) Following export to the cytoplasm, the pre-miRNA is further processed by Dicer removing
the hairpin loop and generating a ~22 nt RNA duplex. 5) Strand selection takes place upon RISC
loading where the selected guide strand directs RISC mRNA targeting. 6) Translational repression
or 7) mRNA degradation takes place based on the level of miRNA: mRNA complementarity and the
specific Argonaute loaded in the RISC. Figure sourced from (Barron et al. 2011d).
31
In order to achieve target gene binding, the pre-miRNA product is further processed by a
second RNase III type enzyme, Dicer, which partners with a double-stranded RNA-
binding protein TRBP (TAR RNA binding protein) to remove the loop generating a ~22
nt RNA duplex (Koscianska, Starega-Roslan & Krzyzosiak 2011), (Fig. 1.1 4). The
Dicer-TRBP complex closely associates with a member of the Argonaute family (Ago2)
that mediates gene silencing and is intimately involved in the loading of the mature
miRNA strand into the (RISC) RNA-Induced Silencing Complex (Chendrimada et al.
2005) (Fig.1.1 5). Subsequent to physical association of the miRNA duplex with all Ago
proteins, the rate limiting step in miRNA mediated gene inhibition, is in the conversion
from inactive RISC complex (Ago associated with RNA duplex) to active RISC complex
(Ago associated with ssRNA) with thermodynamic stability of the duplex having the
greatest influence (Gu et al. 2011). This thermodynamic stability in the 5’ end of the
duplex tips the balance of strand selection (Schwarz et al. 2003, Khvorova, Reynolds &
Jayasena 2003) with as little a single mismatch in the first four base pairs being enough
to favour guide strand selection (Hu et al. 2009). Upon strand selection, the passenger
strand (eg. miRNA-34a*) is rapidly degraded. Passenger strand protection from
degradation has been observed in species such as C. elegans (Chatterjee et al. 2011). The
means by which a passenger strand that is not thermodynamically favourable is selected
is still an unknown process but is likely dependent on co-factors associated with the RISC
complex. Sporadic selection of the passenger strand completely changes the dynamics of
miRNA directed gene inhibition by introducing a sequence that will target a separate set
of mRNAs.
1.2.3 Genomic Organisation of microRNA and Nomenclature
The transcribed primary (pri) miRNA is between 3 and 4kb, and is structurally equivalent
to that of a transcribed mRNA with flanking 5’ capped and 3’ polyadenylation consensus
sequences forming defined boundaries (Cai, Hagedorn & Cullen 2004, Saini, Griffiths-
Jones & Enright 2007). Approximately 50% of miRNAs are expressed from the introns
of protein coding transcripts while the remaining are located within intergenic regions
(Figure 1.2). The latter intergenic miRNAs are expressed independently from their own
unique upstream promoter often located within flanking genes coding sequences.
However, intragenic miRNAs, both intronic and exonic (Olena, Patton 2010), are
32
commonly co-expressed with their host gene and enter the miRNA processing pathway
following intron splicing. The host gene/miRNA co-expression network has implicated
these miRNAs in feedback loops with the host gene itself on occasion being the target
(Ronchetti et al. 2008, Dill et al. 2012). The miRNA cluster miR-17-92 is an example of
an intronic miRNA under the transcriptional control of its own dedicated transcriptional
unit (TU) (Ota et al. 2004a).
Figure 1.2: Genomic Organisation of microRNA
Figure 1.2: Genomic organisation of microRNA. There are four ways in which miRNA can organise
within the genome. 1) Intergenic miRNA are located within the non-coding “junk” DNA between
coding genes, are under the transcriptional control of their own TU and can be 1a) monocistronic or
1b) polycistronic. 2) Intronic miRNA are processed from the intron of protein coding genes
subsequent to intron splicing and can be 2a) monocistronic, 2b) polycistronic and uncommonly 2c)
miRtronic whereby the mature pre-miRNA constitutes the entire intron and by-passes cleavage by
Drosha/DSCR8 through intronic splicing. All intronic miRNA can be expressed under the control of
their own transcription units (TUs) or be co-expressed with their host gene under its own TU. 3)
Exonic miRNA are located within the coding exon of a host gene and can be under either its own TU
or the host genes TU. This figure was sourced from Kelly and colleagues (Manuscript accepted)
33
As of 2008 between 36% and 45% of miRNAs were found to be located in clusters in
mouse and human (Griffiths-Jones et al. 2008). miRNA clusters are transcribed as a long
pri-miRNA transcript and cleaved into several pre-miRNAs by the Microprocessor.
miRNA clusters containing homologous and non-homologous miRNAs can result in
multiple targeting of a particular mRNA target by several miRNAs (Smith et al. 2012) or
the co-suppression of multiple mRNA targets (Kim et al. 2009b). The largest miRNA
cluster identified to date is C19MC which consists of 46 pre-miRNAs expressed in
tandem from a 100kb primary transcript (Bortolin-Cavaille et al. 2009). An intragenic
miRNA of interest is the “mirtron” by which the genomic organisation is unique to any
other type of intronic miRNA. In this instance the full mature pre-miRNA constitutes an
entire intron between two flanking exons whereby intron splicing serves the role of pre-
miRNA cleavage usually filled by DSCR8/Drosha yielding a mature pre-miRNA with
the appropriate secondary structure for recognition and removal from the nucleus by
Exportin-5 (Okamura et al. 2007, Sibley et al. 2012).
After the discovery of lin-4 and let-7 as a class of small temporal regulatory RNAs, the
widespread identification of “microRNAs” in both animals and plants shortly followed
with strong homology across species (Lee, Ambros 2001, Lagos-Quintana et al. 2001,
Lau et al. 2001). Since the formation of the miRBase database in 2002 over 17,000 mature
miRNA sequences in over 140 species have been registered (Griffiths-Jones 2004,
Kozomara, Griffiths-Jones 2011). With the high throughput identification of novel
miRNAs, it was vital for a nomenclature to be developed that catalogued this diverse
family of small RNAs. The first criterion in the miRBase catalogue of miRNA
classification is the species affiliation denoted by a three-four letter prefix of the species
name. For example, hsa-miR-17 indicates that this miRNA is found in human (Homo
sapiens) or CHO (Cricetulus Griceus). Additionally, “miR” designates the mature form
of the miRNA, that being fully processed and RISC loaded, in contrast to “mir” which
indicates the pri-miRNA or pre-miRNA sequence. miRNAs can give rise to two
individual mature miRNA that are derived from the same pre-miRNA based on
thermodynamic stability of the duplex ends, miRNA-17-5p and miRNA-17-3p are
selected from the 5’ and 3’ ends of the hairpin loop respectively. Designation of the
mature miRNAs as -5p or -3p over miR/miR* arose from the lack of experimental data
indicating which arm of the miRNA duplex was predominantly selected, therefore the
miRNA* species could have been originally named -5p or -3p (Ambros et al. 2003).
34
Considering their small size, a high degree of similarities between unique miRNAs is
common and for this reason homologous miRNAs are separated into groups. miRNA-
23b and miRNA-23a differ in sequence by a single nucleotide. Given that this nucleotide
difference may not necessarily be within the seed region, these miRNAs could target
similar genes, placing them in the same “seed family”. As the evolution of miRNA
families can be attributed to duplication events (Sun et al. 2012, Ehrenreich, Purugganan
2008) and subsequent mutagenesis (Lee et al. 2011), a system has been devised that infers
that two miRNAs are identical, paralogous, but centred in separate chromosomal
locations (miR-92a-1 and miR-92a-2). The term “isomiR” refers to the identification of
homologous miRNAs from the same sample source that differs purely by the addition of
a nucleotide at either end of the mature sequence. This discrepancy in miRNA sequence
length can be attributed to the variability in Drosha or Dicer cleavage position during
processing of the pre-miR hairpin (Morin et al. 2008). However the Argonaute protein
Ago-2 has been hypothesised to be involved in the generation of isomiRs through
trimming of the 3’-terminal end of the bound miRNA (Juvvuna et al. 2012).
1.2.4 microRNA mode of action
The catalytic engine driving miRNA-mediated RNA interference is the RISC (RNA-
Induced Silencing Complex) which is composed of two vital components, the single-
stranded mature miRNA that directs RISC targeting and a member of the Argonaute
family of proteins. Mammals encode four Argonaute paralogs referred to as Ago1-4 with
characterisation of this family being based on the presence of the N, PAZ, MID and PIWI
domains (Boland et al. 2011, Ma, Ye & Patel 2004). Despite the overlapping participation
of the four Argonaute paralogs, Ago-2 has been implicated as the main driving force
behind miRNA-targeted mRNA cleavage (Liu et al. 2004, Meister et al. 2004) miRNAs
regulate the transcription of their target mRNAs through base pairing of the “seed” region
situated at 2-8 nt from the 5’ end (Chi, Hannon & Darnell 2012). High complementarity
between miRNA: mRNA is more commonly associated with target mRNA cleavage, as
seen with siRNA, with the more common base-pair mismatching resulting in translational
inhibition. As mentioned earlier, through the generation of IsomiRs, this degree of
miRNA: mRNA complementarity can hypothetically be optimized in some cases through
Ago-2 mediated 3’ terminal trimming of the bound mature miRNA (Juvvuna et al. 2012,
35
Han et al. 2011). Although both the 5’ and 3’ ends of miRNAs can vary widely depending
on the Drosha/Dicer-dependant addition or deletion of 1-2 nt, alterations in the 5’
sequence has the potential to alter the seed sequence greatly resulting in changes in
miRNA targeting specificity (Krol, Loedige & Filipowicz 2010). Activated RISC
complexes degrade target mRNAs or inhibit translation through target binding with
complementary/semi complementary sequences within the 3’UTR with binding sites
often being found within the 5’ UTR and open reading frame (Kloosterman et al. 2004,
Lytle, Yario & Steitz 2007). miRNA recognition elements (MREs) within the 3’UTR are
further underlined by the observation that these functional target sites appear to be under
evolutionary conservation (Kertesz et al. 2007). The RISC complex has also been
demonstrated to mediate specific cleavage at the sarcin-ricin loop (SRL) of 28S ribosomal
RNA eliciting translational suppression by direct ribosomal interaction (Pichinuk, Broday
& Wreschner 2011). Accelerated mRNA decay through miRNA directed de-adenylation
has also been observed (Djuranovic, Nahvi & Green 2012, Mishima et al. 2012).
Despite miRNAs being associated with the fine-tuning of gene expression through
inhibitory affects, miRNA-10a has been demonstrated to enhance the translation of
ribosomal proteins through interactions with their 5’ UTR (Orom, Nielsen & Lund 2008).
1.2.5 Target prediction algorithms and online databases/tools
The potential for an individual miRNA to target a large number of protein coding mRNAs
has its advantages in relation to cellular homeostasis and as possible targets for genetic
manipulation but with this carries disadvantages to the researcher. With the small size of
the seed region being of 6-8 nt in length and the minimum of a 6 bp match causing mRNA
translational suppression (Brennecke et al. 2005), the challenge of predicting mRNA
targets with a high degree confidence poses difficult without experimental validation.
PicTar (Krek et al. 2005), miRanda (John et al. 2004) and TargetScan (Lewis et al. 2003)
are three algorithms, among others, that utilize the classical method of searching for
miRNA targets based on the near complementarity between the 5’ miRNA seed region
and the target mRNAs 3’UTR; conservation of binding sites across species and binding
energy of the miRNA target duplex. Other prediction algorithms expand on the criteria
listed above including the incorporation of G: U wobble pairs within the seed region,
target site accessibility and local concentration of redundant miRNA patterns with a
36
degree of flexibility in computer learning. Variation in the binding criterion applied to
target prediction across the various databases is shown in table 1.4. Although these in
silico tools can identify real biological targets, the researcher is presented with a
substantial list of predicted target genes that must be experimentally verified.
Table 1.4: Summary of the different software algorithms available for miRNA target
prediction
Table 1.4: The prediction algorithms available and the binding criteria utilised in miRNA target
prediction. This table was sourced from a review by Barron et al (Barron et al. 2011d) extensively
assessing the role of miRNAs as candidates for CHO cell engineering.
Furthermore, most prediction algorithms rely heavily on the central “seed” dogma of
miRNA-target binding (all miRNAs contain a 5’-seed region) which is not always the
case (Lal et al. 2009a). Reducing the number of false-positives is an essential feature for
37
novel prediction algorithms. miRGator encompasses the predictive capacity of three well
known algorithms (miRanda, PicTar and TargetScanS) providing information on
expression, functional annotation, pathway and disease (Nam et al. 2008) with an intimate
link to the Tarbase database for experimentally tested miRNAs across 8 organisms
(Sethupathy, Corda & Hatzigeorgiou 2006, Vergoulis et al. 2012). Another online
database, miRWalk, expands the binding criterion outside that of the 3’UTR
encompassing the 5’-UTR promoter region (Lytle, Yario & Steitz 2007) and nucleotide
coding sequence (Tay et al. 2008) spanning three species (human, mouse and rat)
including both predicted, overlapping with predicted binding sites from 8 other prediction
algorithms, and validated miRNA information sourced from the Pubmed online database
(Dweep et al. 2011a).
The reliability of combining multiple prediction algorithms appears questionable over
one single “accurate” algorithm, reviewed in (Alexiou et al. 2009), however further
experimental undertakings will provide insights into mechanism of miRNA:mRNA target
interactions. One such algorithm, mirWIP, utilises experimental data in order to define
the prediction rules (Hammell et al. 2008) whereas an example of an independent
approach establishes an in silico prediction model based on an integrative “Omics” study
of the mammalian inner ear (Elkan-Miller et al. 2011). Here, while studying both mRNA
and protein levels coupled with a differential expression miRNA profile, subsequent
crossover with prediction algorithms, including microRNA:Target pathway databases
(Jiang et al. 2009, Ruepp et al. 2010), will permit the reduction of false positives allowing
a high confident selection of candidate miRNAs and again enhancing the importance of
experimental observation in order to better understand the miRNA target binding
etiquette.
1.2.6 MicroRNAs in Chinese hamster ovary cells – A role in the bioprocess
The CHO cell is subjected to a number of bio-molecular changes during the bioprocess
such as mechanical stress, metabolite accumulation, hypoxia, nutrient limitation,
osmolality, etc., which dampen the capacity for optimal protein production. The use of
RNA interference (RNAi) technology as an application to improve CHO cell specific
productivity and recombinant protein quality ushered in the exploration of miRNAs as an
alternative engineering strategy to target multiple genes in a single pathway or several
38
molecular pathways without impacting on the CHO cells translational machinery,
reviewed in (Wu 2009).(Wu 2009) The identification of miRNAs in the pathology of
diseases such as cancer through either their deletion or up-regulation implicated them as
primary targets for intervention. This stimulated the interest in manipulating the CHO cell
phenotype using miRNA.
The pioneering study by (Gammell et al. 2007) explored potential miRNAs involved in
growth, increased productivity and sustained viability in CHO-K1 during a temperature
shift to 31˚C, a commonly exercised protocol in order to extend the CHO production
phase (Kumar, Gammell & Clynes 2007) identifying a novel CHO miRNA involved in
growth, cgr-miR-21. The exploitation of next-generation sequencing (NGS) technologies
has led to the identification of over ~400 conserved mature miRNA sequences with 387
being identified in a single undertaking using small RNAs from 6 biotechnologically
relevant CHO cell lines (Hackl et al. 2011) with a similar study identifying a plethora of
miR:miR* homologs in CHO across several species using illumine sequencing (Johnson
et al. 2011). Upon the release of the CHO-K1 genome sequence (Xu et al. 2011)
advancements in CHO genetics such as the identification of the genomic loci for over
~200 miRNA through computational analysis (Hackl et al. 2012) will allow for the
identification of novel CHO promoters of miRNAs that are temporally stimulated during
the bioprocess or for the site specific overexpression or deletion of endogenous miRNAs.
Incorporation of this newly acquired genomic sequence data has been compiled into an
online data base (www.CHOgenome.org) (Hammond et al. 2012).
Differential miRNA expression profiles have been carried out on CHO cells in order to
better understand the miRNA dynamics involved in numerous industrial process such as
engineered CHO cells (Lin et al. 2010) and CHO cells at different time points (Hernandez
Bort et al. 2012) including clones with various growth rates (Clarke et al. 2012). A proof-
of-concept study based on a temperature shift differential expression profile, implicated
the over-expression of the miRNA, miR-7, to have a negative effect on cell growth and a
positive effect on cell specific productivity with no influence on apoptosis in a CHO-
SEAP cell line (Barron et al. 2011a). Proteomic analysis of miR-7 over-expression
demonstrated an overlap in proteins up-regulated or down-regulated with the pathways
that were seen to be manipulated (Meleady et al. 2012a).
39
Some of the cellular processes central to the bioprocess have been shown to be impacted
by miRNA regulation in other cell types and turn out to be conserved in CHO, such as
the impact of miR-7 deregulation (Barron et al. 2011b), whereas others have not been
described elsewhere, as in the case of miR-466h and its role in inhibiting apoptosis in late
stage CHO culture (Druz et al. 2013). Stable depletion of miR-7 using a miRNA sponge
vector improved peak cell density, prolonged culture longevity and boosted secreted
protein yield by almost 2-fold in a fed-batch process (Sanchez et al. 2013b). Similarly,
Jadhav and colleagues enhanced the overall yield of an Epo-Fc protein through the stable
over-expression of miR-17 (Jadhav et al. 2014). Both studies demonstrate the utility of
miRNA-mediated engineering to improve(Lai, Yang & Ng 2013b) CHO cell growth
without negatively impacting on specific productivity, a common trade-off for superior
growth rates and vice versa (Lai, Yang & Ng 2013b). From a product quality perspective,
Strotbek et al., reported that the glycosylation profile of recombinant IgG1 was not
compromised in CHO cells engineered to stably co-express miR-557 and miR-1287,
while inducing increased specific cellular productivity (Strotbek et al. 2013). Besides the
manipulation of individual miRNAs, it has also been shown that tuning global, cellular
miRNA levels via the RNAi processing machinery influences bioprocess-relevant
phenotypes. Hackl et a., (2011) found that overall miRNA expression in CHO cells
cultured in serum-containing medium was elevated (Hackl et al. 2011) and upon further
investigation found that Dicer mRNA and protein levels decreased in response to nutrient
or serum removal. However, attempts to engineer Dicer expression demonstrated a very
sensitive phenotypic response to the levels of Dicer present in the cells. Knockdown
below endogenous levels or excessive over-expression both negatively affected growth
but mild overexpression was found to be beneficial (Hackl et al. 2014b). For this reason,
fine tuning the expression of single or small numbers of individual miRNAs, such as those
mentioned previously, is probably a better approach for those interested in targeting a
specific cellular function or behaviour. A number of groups have reported on miRNA
expression profiling in CHO cells representing different phenotypes, producing different
products or grown under different conditions in order to identify potential engineering
targets.
Striving to improve the performance of CHO cells in industry has led to the production
of hundreds of high quality biopharmaceuticals, sequencing of the CHO genome and now
the identification of hundreds of microRNAs that are highly conserved and endogenously
40
encoded. Efforts to optimise the CHO phenotype using miRNAs are proving promising
as an alternative to gene based engineering by achieving the same phenotypic outcome
with a lightened dependency on the cells own translational machinery. Some well-known
and routinely characterised miRNA families represent a major outlet for CHO cell
engineering when considering the conserved nature of their mechanistic actions across
numerous species and disease states, such as cancer, with one family in particular, miR-
34, representing the first miRNA destined to take part in clinical trials (Bader 2012). Two
miRNA clusters are described below that exhibit considerable interest for CHO cell
engineering in addition to the wealth of candidates that offer viable engineering targets
(see table 1.5)
Table 1.5: Candidate miRNA for CHO cell engineering
Biological
process
microRNA Cluster Phenotype Target Reference
Cell cycle
miR-7 (CHO) No Tumour
suppressor
Skp2
Psme3
(Sanchez et al.
2013a)
miR-34a No Tumour
suppressor
CCND1,
CCNE2,
CDK4/6
(Sun et al.
2008a, He et al.
2007b)
miR-24 miR-23a~24-
2
miR-23b~24-
1
Tumour
suppressor
E2F2
MYC
(Lal et al.
2009b)
miR-17
(CHO)
miR-17~92 Oncogenic CDKN1a/p21 (Ivanovska et
al. 2008)
miR-31 No Oncogenic LATS2
PPP2R2A
(Liu et al.
2010b)
Apoptosis
Mmu-miR-
466h
(CHO)
mmu-mir-
297-669
Pro-apoptotic Bcl212
Birc6
(Druz et al.
2013)
miR-34a No Pro-apoptotic Bcl2,
surviving
(Bommer et al.
2007, Shen et
al. 2012)
miR-15a/-16 mir-15a~16-1 Pro-apoptotic PDCD4 (Calin et al.
2008)
miR-21 No Pro-apoptotic HSP60
HSP70
(Itani et al.
2012)
miR-133 miR-1/-133 Pro-apoptotic Caspase 9 (Xu et al. 2007)
Culture Stress
miR-126 No Shear Stress Bcl2,
FOXO3, Irs1
(Zhou et al.
2013)
miR-31 No Hypoxia HIF (Liu et al.
2010)
miR-429 miR-
200b~429
Osmolarity
responsive
OSTF1 (Yan et al.
2012)
miR-375 No Lactate
production
LDHB (Kinoshita et al.
2012)
41
miR-144 miR-
4732~451a
Oxidative stress NRF2 (Sangokoya,
Telen & Chi
2010)
Energy
Metabolism
miR-23a miR-23a~24-
1
Glutamine
metabolism
GLS (Gao et al.
2009)
miR-124 No Glycolysis PKM1/2 (Sun et al.
2012b)
Let-7 Let-7a/d/f Glucose
metabolism
Irs2, INSR (Frost, Olson
2011)
miR-23 miR-23a~24-
1
Mitochondrial
biogenesis
TFAM (Jiang et al.
2013b)
miR-122 miR-
122~3591
Lipid
metabolism
Cholesterol
related genes
(Esau et al.
2006b)
Protein
Quality
miR-34a No Core
fucosylation
FUT8 (Bernardi et al.
2013b)
miR-148b No N-Glycosylation C1GALT1 (Serino et al.
2012)
miR-30b/d miR-30b/d O-Glycosylation GALNT7 (Gaziel-Sovran,
Hernando
2012)
Protein
productivity/
secretion
miR-33a/33b No GSIS Irs2 (Davalos et al.
2011)
miR-30c* No UPR XBP-1 (Byrd, Aragon
& Brewer
2012)
miR-410 mir-
323b~656
Protein
secretion
? (Hennessy et al.
2010)
miR-34a No Insulin secretion VAMP2 (Lovis et al.
2008)
miR-124 No Insulin
exocytosis
SNAP25
synapsin-1A
(Lovis,
Gattesco &
Regazzi 2008) Abbreviations: Bcl2-B-cell lymphoma 2, Birc6-Baculoviral AIP containing repeat 6, C1GALT1- core 1 synthase,
glycoprotein-N-acetylgalactosimine 3-beta-galactosyltransferase, 1, CCND1/E2-Cyclin D1/E2, CDK4/6-Cyclin-dependent
kinase 4/6, CDKN1a-cyclin-dependent kinase inhibitor 1a, FOXO3-Forkhead box O3, GALNT7- N-
acetylgalactosaminyltransferase 7, GLS-Glutaminase, GSIS-Glucose stimulated insulin secretion, Heat-shock protein-
60/70, HIF-Hypoxia inducible factor, Irs2-Insulain receptor substrate 1/2, INSR-Insulin receptor, LDHB-Lactate
dehydrogenase B, LATS2-Large tumour suppressor 2, NRF2-Nuclear factor erythroid 2-related factor 2, OSTF1-
Osteoclast stimulating factor 1, PPP2R2A-PP2A regulatory subunit B alpha isoform, PDCD4-Programmed cell death 4,
PKM1/2-Pyruvate kinase 1/2, Psme3-Proteomsome activator subunit 3, Skp2-S-phase kinase associated protein 2,
SNAP25-synaptosomal-associated protein 25, TFAM-Mitochondrial transcription factor A, UPR-Unfolded protein
response, VAMP2-Vesicular associated membrane protein 2, XBP-1-X-box binding protein-1.
1.2.6.1 miR-17~92 cluster
One of the best characterised oncogenic miRNAs is miR-17~92, a polycistronic miRNA
cluster, also known as oncomiR-1, that is among the most potent oncogenic miRNA. The
precursor transcript for this miRNA contains 6 stem-loop hairpin structures in tandem
that ultimately generate 6 mature miRNAs: miR-17, miR-18a, miR-19a, miR-20a, miR-
19b-1 and miR-92-1. Human miR-17~92 is located at 13q31.3, a region of the genome
that is frequently amplified in several hematopoietic malignancies and solid tumours,
such as diffuse B-cell lymphomas (DLBCLs), follicular lymphomas, Burkitt’s
lymphomas and lung cancer (Ota et al. 2004b). Studies have revealed that miR-17~92
42
acts pleiotropically during both normal development and malignant transformation to
promote phenotypes such as proliferation, inhibit differentiation, augment angiogenesis
and sustain cell survival (Fig 1.3). The anti-apoptotic nature of the miR-17~92 cluster is
evident in progenitor B-cells whereby c-myc induced apoptosis is suppressed by the
transcriptional activation of this cluster (Tagawa, Seto 2005). Furthermore, the oncogenic
transcription factor family E2F (E2F1/E2F3), whose transcription drives the cell through
to S phase through the activation of S phase genes such as thymidine kinase, DNA
polymerase and Cyclins E and A, has been demonstrated to be a direct transcriptional
activator of miR-17~92 (Sylvestre et al. 2007). Similar to c-myc, the overexpression of
E2F1 in normal and malignant cell types is sufficient to induce p53-mediated cell cycle
arrest and apoptosis (Hsieh et al. 1997).
Figure 1.3: Pleiotropic functions of miR-17~92
Figure 1.3: Above demonstrates the Pleiotropic functionality of the miR-17~92 cluster and its
component members to generate and sustain a more aggressive cancer phenotype. Figure was
sourced from the review by V. Olive and colleagues, 2010 (Olive, Jiang & He 2010b).
As E2F1 acts as a transcription factor for itself it is caught in a perpetual positive feedback
loop. However, through transcriptional activation of the miR-17~92 cluster, not only is
fine-tuning of its own translation controlled by miR-17-5p (O'Donnell et al. 2005a) but
entrance into apoptosis is additionally controlled through the interaction of the miRNA
cluster with apoptotic components such as PTEN and Bim (Ventura et al. 2008).
Tumourigenicity and aggressiveness has further been demonstrated by Dews and
colleagues in colonocytes through the promotion of angiogenesis by targeting and
43
repressing anti-angiogenic factors such as thrombospondin-1 (TPS-1) and connective
tissue growth factor (CTGF) by miR-18a and miR-19 respectively (Dews et al. 2006a).
Given the wide characterisation of this miRNA cluster, it is not surprising that it has
shown up on the radar in miRNA research in CHO. Two independent studies revealed the
potential of the miR-17~92 cluster as an engineering tool for the manipulation of CHO
cell growth. In the first instance, next-generation sequencing unveiled the presence of 387
mature miRNA transcripts in CHO with the members of the miR-17~92 cluster being
among them. Furthermore, cDNA sequencing of 26 previously validated targets (E2F1,
CCND1, BCL211, MAPK14, PTEN, RB1 and VEGFA) of this cluster demonstrated seed
target site conservation suggesting conserved functionality of miR-17~92 in CHO (Hackl
et al. 2011). Transient overexpression of miR-17 was carried out by the same group to
validate its impact on the proliferation profile of CHO. An increase of 15.4% compared
to negative controls was reported for this miRNA suggesting the utility of the entire
cluster as an engineering tool as being functionally viable (Jadhav et al. 2012). Secondly,
Clarke and colleagues identified the miR-17~92 as being differentially expressed in fast
growing CHO cells with each of the cluster members exhibiting a positive correlation in
clones with increasing growth rates (Clarke et al. 2012).
1.2.6.2 miR-23~27~24 clusters
In humans, the miR-23a~27a~24-2 cluster is intergenic and localised on chromosome
19p13 while its paralog miR-23b~27b~24-1 is an intronic cluster (Aminopeptidase O,
AMPO (Sikand, Slane & Shukla 2009)) localised to chromosome 9q22 (Chhabra, Dubey
& Saini 2010a). The deregulation and aberrant expression of both of these miRNA
clusters have been associated with numerous cancerous phenotypes such as miR-
23b~27b~24-1 in prostate cancer (Ishteiwy et al. 2012) and miR-23a~27a~24-2 in gastric
cancer cells (An et al. 2013). Individually however, each member of these clusters has
been associated with numerous cellular phenotypes including cellular proliferation, cell
cycle, apoptosis, DNA repair and cellular metabolism. In primary keratinocytes, the
overexpression of miR-24 was demonstrated to target the cyclin-dependent kinase
inhibitors (CDKIs) p27Kip1 and p16Ink4a thus promoting cellular proliferation (Giglio et al.
2013). In contradiction to this mechanism of action, the targeting of key nodes of the cell
cycle by miR-24 triggered cell cycle arrest in G1 phase through reported gene targets such
44
as MYC and E2F2, as well as their downstream targets CCNB1, CDC2 and CDK4 (Lal et
al. 2009a). Furthermore, miR-24 levels were determined to be, in one study, highly
expressed in patient breast carcinoma samples compared to benign breast tissues. Ectopic
overexpression of miR-24 promoted breast cancer cell invasion and migration while
overexpression of its direct targets PTPN9 and PTPRF reversed this observed phenotype
(Du et al. 2013).
From a metabolic perspective, recent reports have implicated c-Myc to have a role in the
catabolism of glutamine through the repression of miR-23a and miR-23b (Gao et al.
2012)(Dang 2010), thereby integrating this transcription factors role in cellular
proliferation and metabolism. Altered glucose metabolism, also known as the “Warburg
effect”, is a well-studied metabolic phenotype of cancer. This mechanism of “aerobic
glycolysis” is characterised by the excessive uptake of glucose and conversion to lactate
despite the availability of sufficient oxygen levels (Bensinger, Christofk 2012). C-Myc
also collaborates with hypoxia-inducible factor-1 (HIF-1) to alter mitochondrial
respiration resulting in the increased rate of glycolysis sufficient for tumour cell
adaptation to its hypoxic (oxygen deficient) microenvironment, providing a growth
advantage through the production of ATP and NADPH via excessive turnover of glucose
and glutamine (Dang 2010). The link between glutamine catabolism and the increase in
the enzyme glutaminase (GLS) has been associated with the repression of miR-23a/b,
direct targets of GLS, by c-Myc in neoplastic cells (Gao et al. 2009).
The third member of these two clusters miR-27a/b has been associated with cellular
metabolism through adipogenesis regulation via interference of the key regulator PPARγ
(Lin et al. 2009). Overexpression of miR-27a/b was demonstrated to promote endothelial
cell sprouting through targeting of the angiogenesis inhibitor semaphoring 6A (SEMA6A)
(Urbich et al. 2012). Not only has the oncogenic miR-27 been associated as an interfering
factor in the G2/M check point of the cell cycle in breast cancer (Mertens-Talcott et al.
2007) but it has additional clinical value as a prognostic marker for breast cancer
progression and patient survival with anti-correlated expression with its target ZBTB10
(Tang et al. 2012).
It can be seen that the component members of the miR-23~27~24 clusters are involved
in multiple cellular phenotypes from the opportunistic utilisation of carbon sources to
direct activation or inhibition of pro-/anti- cell cycle regulators. Each member has
45
antagonistic to synergistic functionality, targeting multiple different aspects of cellular
homeostasis thereby generating a tailor designed, and thriving cancer cell. The part they
play in key cellular processes leading to tumorigenesis could also be potentially translated
to CHO cell engineering.
1.3 microRNAs: Tools for CHO cell engineering
Various tools exist for the manipulation of miRNAs suitable for CHO cell engineering,
reviewed in (Barron et al. 2011e). Although, upon publication of this review, miRNAs in
CHO was still in its infancy, various applications have been published from transient to
stable miRNA manipulation on a wealth of recombinant CHO cells with a great degree
of success, reviewed in (Jadhav et al. 2013).
1.3.1 microRNA identification
Numerous differential expression profiles have been reported to date in CHO cells under
various culture conditions such as temperature shift (Gammell et al. 2007, Barron et al.
2011b), nutrient depletion (Druz et al. 2011), growth rate profiling (Clarke et al. 2012)
and culture stage profiling (Hernandez Bort et al. 2012). The conservation of mature
miRNAs has allowed the use of human miRNA arrays to be used for the identification
global miRNA changes in CHO cells including mouse and rat hybridization. Northern
blotting has been used for the identification of pre-miRNA and mature miRNA through
the use of short sequence-specific probes (Chen et al. 2004). More recently however,
individual miRNAs can be detected based on their mature processed sequence by
specifically designed reverse transcription (RT) primers containing a 5’-loop structure.
This quantitative RT-PCR (qPCR) method can be SYBR-Green or Taqman based. qPCR
based applications can also be used for more “global” large-scale miRNA profiling using
384-well arrays, each containing single miRNA probes, which gives a significant and
more comprehensive snapshot of miRNA interaction simultaneously (Weber et al. 2010).
The use of hybridisation arrays, miRNA bio-arrays generated with probes against human,
rat and mouse miRNAs, have been exploited to identify expression patterns in a similar
way to conventional transcriptional profiling methods. This approach identified 26
differentially expressed miRNAs in CHO subject to temperature shift and was the first
miRNA profiling study carried out in this cell type (Gammell et al. 2007). With the release
46
of the CHO-K1 genome and the addition of over 400 mature CHO miRNA sequences to
the miRBase database, it is only a matter of time until CHO-specific arrays are developed
for the global characterisation of miRNAs intimately involved within different cellular
processes of biotechnological interest.
1.3.2 Transient microRNA interference
To verify the utility of predicted miRNAs as engineering tools, phenotypic assessment is
performed through the increase or decrease of endogenous miRNA levels. In the case of
gain-of-function, mimics take the form of 22 nt dsRNA duplexes that are similar in
structure to that of pre-processed or “diced” miRNAs (Fig. 1.1 4) (Ford 2006). This
introduction of pre-processed miRNAs circumvents the requirement of the utilisation of
the primary miRNA processing machinery such as Drosha/DGCR8 and Exportin-5
providing immediate activity with lipid-based delivery remaining the primary method of
transient introduction (Tseng, Mozumdar & Huang 2009). These mimics take the form of
short-hairpin (sh) RNAs and do not possess the base-pair mismatches associated with
natural occurring miRNA duplexes. The benefit of this is that miR* strands can be
efficiently overexpressed and analysed, despite their “normal” endogenous structures
favouring selection of the guide due to thermodynamic properties (Schwarz et al. 2003).
Transient miRNA intervention using mimics has been reported in CHO in the case of
miR-7 inducing a temperature-shift like phenotype (Barron et al. 2011b) and the library
screen of 879 miRNAs assessing their role in productivity (Strotbek et al. 2013).
With the identification of miRNAs being frequently located at fragile chromosomal
locations susceptible to genomic instability resulting in chromosomal deletions or
transcriptional attenuation (Lagana et al. 2010) it was important to develop an efficient
method to evaluate miRNA loss-of-function phenotypes and this has subsequently been
successfully employed through numerous means, reviewed in (Stenvang et al. 2012).
Chemically modified antisense oligonucleotides (AON), antimiRs, are a widely
employed approach to accomplish miRNA loss-of-function by sequestering the mature
target miRNA, by strong hybridisation, inherently in competition with cellular mRNA
targets resulting in translational de-repression (Dean 2001). The chemical modification
of these small RNAs through various methods has been executed in order to prolong the
stability and subsequently their efficacy within the cellular environment, reviewed in
47
(Stenvang et al. 2012). These modifications include the 2-OH residue of the ribose of the
antisense RNA being replaced with 2’-O-methyl (Panzitt et al. 2007), 2’-O-methoxyethyl
(Esau et al. 2006a), 2’-Fluoro (Lennox, Behlke 2010) and locked nucleic acid (LNA)
(Chan, Krichevsky & Kosik 2005). Modifications confer enhanced benefits such as
resistance to nucleases, prolong stability through stable hybridization, increase binding
affinity without abrogating specificity and enhance melting temperature (Tm). A modified
version of LNAs, termed 8-mer LNA-antimiRs, has been developed to assess the
biological functions of entire miRNA seed families (Obad et al. 2011). To reduce off-
target effects such as inhibition of entire seed families, RNAi constructs have been
designed specific to the primary miRNA sequence allowing a higher degree of specificity
(Vaistij et al. 2010). Transient miRNA inhibition is routinely used in CHO cell screening
such as in the case of miR-7 (Barron et al. 2011b) and miR-466h (Druz et al. 2011).
The large-scale transfection of miRNAs using non-viral delivery has been demonstrated
for mammalian cell cultures such as CHO in complex protein production media (Fischer
et al. 2013). This offers new potential for miRNAs to be used in existing approved
processes via transfection, having less regulatory implications and ultimately avoiding
stable cell line development (Fig. 1.4).
48
Figure 1.4: Transient miRNA intervention within the bioprocess
Figure 1.4: Large-scale bulk transfection can be performed transiently by introducing small working
concentrations of either mimics or inhibitors at various culture time points in a bioreactor to achieve
a multitude of phenotypes such as enhanced quality or yields. This figure was sourced from Kelly et
al., 2014 (Manuscript accepted).
Not only does transient overexpression bypass the natural miRNA processing pathways
but synthetic miRNA stability does not allow the elucidation of prolonged phenotypic
impacts (Thomson, Bracken & Goodall 2011). For this reason and from the perspective
of ultimately engineering CHO cell lines for industrial application, stable methods of
miRNA intervention are required.
1.3.3 Stable microRNA interference
Stable miRNA intervention is required to overcome the inherent pitfalls of transient
miRNA assays such as the short half-life of introduced miRNAs, the reduced cytoplasmic
payload of miRNAs across daughter cells upon cellular division (Song et al. 2003,
Bartlett, Davis 2006), the saturation of miRNA biogenesis machinery and the titrational
influence based on the reciprocal relationship between miRNAs and mRNAs (Pasquinelli
2012). In the case of gain-of-function studies, stable overexpression provides a more
physiological natural state of miR expression, preventing off-target effects (Grimm et al.
2006). Furthermore, stable miRNA intervention has been reported to induce a phenotype
in cases where transient intervention proved unsuccessful such as miR-7 in CHO cells
(Barron et al. 2011b, Sanchez et al. 2013b).
Anti-sense oligonucleotide
Anti-miR
miRNA duplex hairpin
pre-miR
Endo mRNA MRE Poly A
Endo mRNA MRE Poly A
Endo mRNA MRE Poly A
Endo mRNA MRE Poly A
Bulk
Transfection
Bioreactor
Target
Protein
Target
Protein
Enhanced product
Output/quality
49
The miRNA hairpin structure has been exploited in order to develop short-hairpin (sh)
RNA vectors for the validation of gene knockdown studies (Zhu et al. 2007). shRNAs
driven by a polymerase (pol) III promoter such as U6 (Fig. 1.5 A) are modelled after
precursor miRNAs (Rossi 2008) containing a sequence of ~22-25 nt, a short loop region
and the reverse compliment to the 22-25 nt region whereas other systems such as pol II
act more like primary miRNAs in which the shRNA is flanked by natural miRNA
sequences (Fig. 1.5 B), termed shRNA-miRs (Meerbrey et al. 2011, Du et al. 2006). The
natural flanks take the form of miR-30 sequences (125 bases 5’ and 3’) forming the
second generation of shRNAs (Silva et al. 2005) and have been shown to provide a higher
knockdown efficiency (Stegmeier et al. 2005) when driven through a pol II promoter.
shRNA-miRs transcribed in vivo from polymerase II, feed directly into the miRNA
biogenesis pathway usurping the entire miRNA processing machinery whereby by
shRNAs driven by pol III circumvent nuclear processing by Drosha acting like a less
“natural” miRNA system (Lebbink et al. 2011).
The observation that miR-flanking sequences and pol II promoters enhanced miRNA
biogenesis in addition to abrogating it upon sequence mutations (Zeng, Cullen 2003) led
to the final evolutionary step of these vectors to directly use the endogenous miRNA
sequence itself (Fig. 1.5 C). 100 bp 5’- and 3’- flanks were identified to be sufficient to
drive miRNA expression in the case of miR-199a (Furukawa et al. 2011). Multi-
expression vectors of miRNAs from various genomics locations (miR-34a, -34b and -
34c), including 60 bp, flanks have demonstrated the possibility of creating artificial
miRNA clusters (Qiu, Friedman & Liang 2011) presenting a flexible system for “off-the-
shelf” tailored miR construct design.
50
Figure 1.5: Construct of siRNA/miRNA expression vectors
Figure 1.5: A schematic diagram following the evolution of A) first generation shRNA construct
transcribed from a pol III promoter as a hairpin that bypasses Drosha nuclear processing. B)
Modified shRNA (shRNA-miR) to include endogenous miR-30 flanking sequences (5’- and 3’-)
transcribed from pol II to yield a primary miRNA precursor that enters the full biogenesis pathway.
C) Complete endogenous miRNA sequence including flanks cloned from genomic DNA and under
the control again of a pol II promoter.
Stable miRNA expression constructs have also been generated using the miR-155
flanking motifs and have been successfully employed in CHO cell engineering such as
miR-17/-21 (Jadhav et al. 2012) and the miR-17~92 cluster (Jadhav et al. 2014).
Subsequent to the publication of the CHO genome (Xu et al. 2011), pre-miRNA
sequences and their respective genomic locations were made available (Hackl et al. 2012).
From this, the use of endogenous CHO sequences has been demonstrated to be more
efficient when compared to their chimeric counterparts using miR-155 flanks (Klanert et
al. 2013). A 3-fold increase in mature miRNA expression was observed in the case of
(Pol III) U6 Promoter
(Pol II) CMV Promoter miR-30 5’ flanking miR-30 3’ flanking
A) First generation shRNA
B) Second generation shRNA (shRNA-miR)
C) Complete microRNA expression construct
(Pol II) CMV Promoter
51
miR-221/222 and miR-15b/16 clusters when endogenous CHO clusters were stably
expressed. The use of endogenous sequences does prove troublesome when attempting to
overexpress miR* strands due to the thermodynamic favourability towards the guide
strand (Gu et al. 2011). To navigate around this obstacle, shRNA-based designs have been
exploited, eliminating natural flanking sequences which contribute a bias in strand
selection (Yang et al. 2011) and ensuring almost full complementary of the duplex
partner, as reported for miR-146b-3p (Qu et al. 2012).
One method to assess prolonged miRNA inhibition has been the introduction of miRNA-
resistant target genes, circumventing miRNA specific deletion, through site-directed
mutagenesis of the MRE, rendering the miRNA ineffective (Garcia 2008). Global
interference with miRNA expression through genetic deletions of pivotal miRNA
biogenesis machinery such as dicer (Ketting et al. 2001) has also been explored. Stable
knockdown of miRNA through these means, however, proved itself to be challenging and
ineffective as interference with miRNAs during embryonic development has been
observed to abrogate the developmental process (Forstemann et al. 2005). The stable
expression of shRNAs designed specifically to the mature miRNA sequence has been
demonstrated to be effective such as in the case of miR-466h inhibition in CHO cells
ultimately extending culture longevity (Druz et al. 2013). To reduce off target effects,
RNAi constructs can be designed that target the miRNAs primary transcript which would
have a higher degree of uniqueness ultimately allowing for the inhibition of a single
miRNA target and not entire seed families (Vaistij et al. 2010). miRNA sponge
technology offers an attractive option for stable miRNA inhibition as suggested by Barron
and colleagues (Barron et al. 2011e) and has proved its usefulness in the optimisation of
pharmaceutical production cells (Sanchez et al. 2013b).
The increased understanding of miRNAs, their regulatory motifs, their biogenesis and
processing has allowed for the development of numerous expression constructs allowing
the researcher to develop, with minimal understanding, shRNA and miRNA vectors,
making these regulatory molecules easily accessible for industrial applications.
52
1.3.4 microRNA sponge technology
A recent review by Park and colleagues on genetic knockout studies in mice indicated
that “genetic ablation of miRNAs may not result in obvious phenotypes” (Park, Choi &
McManus 2010). This may be attributed in part to the functional redundancy inherent in
the presence of related miRNA seed families as well as paralogs that share endogenous
mRNA targets (Bartel 2009). As such, multiple rounds of genetic knockout would
potentially be necessary to observe any phenotypic impact. Given the genetic complexity
of miRNA organisation, genetic knockouts also pose problematic as the knockout of
intronic miRNAs could interfere with host gene transcription or, in the case of clusters,
interfere with all components instead of one. A means to navigate around this inherent
genetic fortitude has come from observations in nature of the apparent reciprocal
relationship between miRNAs and their target genes, reviewed in (Pasquinelli 2012). This
lead to the hypothesis that miRNA targets could act as competitive inhibitors, competitive
endogenous RNAs (ceRNAs) of miRNA function (Seitz 2009).
The first observation of an endogenous mRNA impacting on miRNA function was
discovered in Arabidopsis thaliana and their response to phosphate starvation (Franco-
Zorrilla et al. 2007). Upon deprivation of inorganic phosphate, up-regulation of miR-399
resulted in the accumulation, rather than depletion, of its predicted target PHO2 (putative-
conjugating enzyme E2). It was discovered that the expected suppression of PHO2 was
circumvented due to increased expression of the non-coding (nc) gene, IPS1. Thus IPS1
acted as a miRNA decoy, sequestering or titrating miR-399 function away from its
endogenous targets. Furthermore, additional mismatches in the duplex formed between
miR-399 and IPS1 prevented Ago2-mediated slicing of the IPS1 transcript, allowing
persistence of this “target mimic” in the cell. The presence of common MREs allow these
ceRNAs to mutually regulate each other in a titration-dependent manner, as observed
recently for the ZEB2 transcription factor and the tumour suppressor PTEN (Karreth et
al. 2011, Salmena et al. 2011). Not only did the observation demonstrate that target
mRNA abundance dilutes miRNA activity, but it identified its presence in a mammalian
system and in contributing to a disease state (Arvey et al. 2010). Indeed this phenomenon
was found to go beyond pairs of protein-encoding genes to include long ncRNAs
(lncRNAs) (Ponting, Oliver & Reik 2009). To support the role of PTEN as a tumour
suppressor, the pseudogene PTEN1 contains MREs within its 3’UTR similar to that of its
53
coding counterpart PTEN. Expression of this pseudogene transcript retains biological
activity as a decoy to sequester and fine tune the translation of the coding PTEN through
competitive binding. Indeed it’s been shown that the PTEN1 locus is frequently lost in
human cancers (Poliseno et al. 2010).
More recently, the phenomenon of head-to-tail splicing of exons has been shown to
generate an exciting new class of non-coding RNAs with regulatory potential, known as
circular RNAs (circRNAs) (Danan et al. 2012, Burd et al. 2010, Hansen et al. 2011). For
example, an antisense transcript from the cerebellar degeneration-related protein 1
(CDR1as) (Memczak et al. 2013) was found to generate a circular decoy that harboured
63 conserved binding sites for the well characterised tumour suppressor miRNA, miR-7
(Zhang et al. 2013). This class of regulatory RNA offers a viable engineering tool due to
its superior stability, potentially as a result of its inherent resistance to cytoplasmic
debranching enzymes and RNA exonucleases (Jeck et al. 2013).
54
Figure 1.6: microRNA sponges divert endogenous miRNA activity
Figure 1.6: microRNA sponges repress microRNA function: A) demonstrates the microRNA
mediated translational repression of the target gene through Watson-Crick imperfect base-pairing
to the miRNA target sequence (Black box) resulting in no protein expression. B) Translational
activation of the endogenous mRNA target through sequestration of the antagonizing miRNA by
decoy miR-sponge sequences (Green box) transiently expressed form a plasmid vector. C) Similar
level of microRNA sponging achieved by a chromosomally integrated sponge containing more MBS
that is expressed at lower levels compared to transient expression from a plasmid in diagram B.
The discovery of these endogenous miRNA decoys led to the development of “microRNA
sponge technology” as an experimental tool to evaluate miRNA function (Ebert, Neilson
& Sharp 2007). Named for their ability to soak up endogenous miRNAs, miRNA sponges
contain multiple sites of complementarity to the miRNA of interest, usually in an artificial
3’UTR placed downstream of a reporter gene such as GFP (Fig. 1.6). Drawing from
lessons in nature, mismatches or “wobble” are introduced as a means to inhibit miRNA-
mediated mRNA cleavage thus prolonging the life-time of the sponge decoy (Kluiver et
al. 2012b). Introduction of “wobble” has been demonstrated to be more effective at
diverting endogenous miRNA function than perfect antisense MREs due to the prolonged
A) mRNA translation repression by miRNA
B) miRNA inhibition through target mimicry: “miR-Sponges”
C) miRNA inhibition through reduced sponge expression with increased MBS
NO PROTEIN PRODUCT
55
life-time of the miR-sponge transcript as a result of miRNA/RISC-mediated mRNA
cleavage inhibtion (Haraguchi, Ozaki & Iba 2009). As seed pairing has been observed to
be sufficient to drive miRNA-target interactions, miR-sponges have the potential to
sequester entire miRNA seed families as demonstrated in the case of the let-7 family
(Yang et al. 2012). Although this can complicate efforts to understand the network of
downstream molecular events that accompany multi-miRNA inhibition, research has
proven the miR-sponge approach to be a viable tool for industrial rCHO cell line
engineering (Sanchez et al. 2013b). In this study, stable diversion of miR-7 through the
presence of a sponge decoy resulted in increased peak cell density, prolonged viability
and ultimately enhanced yield in a CHO fed-batch culture.
Another potential application of this approach is the use of sponge sequences as a means
of controlling the expression of transgenes in a miRNA-dependent manner. For example
if there were a situation where overexpression of a particular gene was desirable only at
a particular point in the culture but otherwise it should be silent. By identifying an
endogenous miRNA whose expression was anti-correlated with the desired transgene
expression profile, a sponge for that miRNA placed downstream of the transgene
sequence would suppress its expression during the chosen culture phase but become de-
repressed later when the miRNA was down-regulated (Fig. 1.7). This application is in
some ways equivalent to the use of inducible promoters.
56
Figure 1.7: Mechanism of exploiting a miRNA-dependent transgene
Figure 1.7: The above schematic diagram demonstrates the hypothetical utility of miRNA sponge
technology when implemented within a CHO cell system of whose miRNA expression profile across
various stages of the culture is known. A plasmid vector expressing the cell cycle inhibitor, p27,
engineered to contain MREs specific for a given miRNA whose expression profile is known across
various phases of CHO cell culture (blue line). In early Exponential growth phase p27 expression (red
line) is maintained at a low level due to the high abundance of mature miRNA specific for the pre-
engineered MREs. Once endogenous miRNA expression is transcriptionally down-regulated, p27
translation gradually becomes de-repressed ultimately halting the cell cycle mimicking a temperature
shift-like phenotype. The above figure was sourced from Kelly and colleagues, 2014 (Manuscript
accepted).
0
20
40
60
80
100
120
0 5 10 15 20
0
20
40
60
80
100
120
0 5 10 15 20
0
20
40
60
80
100
120
0 5 10 15 20
miRNA Exp
p27 Exp
p27 MRE MRE MREMRE p27 MRE MRE MREMRE
p27
Exponential GrowthGrowth Inhibition
Artificial Temp. Shift
Constitutive miR-sensor
expression
0
20
40
60
80
100
120
0 5 10 15 20
miRNA Exp
p27 Exp
Exponential Growth Phase Stationary Growth Phase
CMV p27 MRE PolyA
miRNA-dependent transgene
57
1.4 Glyco-engineering
As of 2012, there were 28 FDA approved recombinant therapeutic antibodies which
represented 36% of the total market revenue for pharmaceutical biologics (Butler,
Meneses-Acosta 2012). The development of chimeric and humanised antibodies marked
a major breakthrough in monoclonal antibody (Mabs) therapy as previous therapeutic
antibodies were of murine origin thus introducing a foreign, highly immunogenic protein
demonstrating suboptimal effector functions with short half-lives (Reichert 2010). The
efficiency of Mabs is due to their antigen specificity and their downstream effector
functions through the formation and activation of immune complexes. All FDA approved
therapeutic Mabs (Table 1.6) are of the immunoglobulin G (IgG) class which in itself is
composed of four subclasses: IgG1, IgG2, IgG3 and IgG4, defined according to their
relative concentrations in normal serum. It has been demonstrated that each subclass
elicits a unique profile of effector functions and that selection of each isotype for clinical
development is dependent on the desired therapeutic outcome and disease type (Jefferis
2007). The utility of monoclonal antibody therapy is layered and when tailor designed in
the right way can achieve several outcomes such as: (i) killing cells or organisms,
cancer/bacteria; (ii) neutralisation of soluble molecules such as cytokines in chronic
diseases or endotoxins; (iii) act as agonists or antagonists of cellular activity; (iv) induce
apoptosis, and (v) delivery of a cytotoxic payload such as a chemotherapeutic drug. The
end point clinical efficacy of therapeutic antibodies is dependent on two mechanisms:
target-specific binding of the Fab (antigen-binding fragment) domain in addition to the
availability of cell surface epitopes and the immune-mediated effector functions
orchestrated through the activation of antibody-dependent cell cytotoxicity (ADCC)
and/or complement-dependent cytotoxicity (CDC) (Chan et al. 2010, Raju 2008).
Immune-mediated effector functions are mediated by interactions between the Fc
(Crystallisable fragment) of the Mab and Fc receptors (FcRs) on various cell types (Siberil
et al. 2007). Although CHO cells are the primary mammalian cell type for recombinant
therapeutic protein production due to their similar posttranslational modifications, it has
now emerged that the variable glycoforms produced in CHO can impose limitations on
effector functions (Nimmerjahn, Ravetch 2008). Such glycoforms can reduce the potency
of ADCC activation, requiring higher therapeutic doses per treatment and ultimately
increase global demand thus placing a strain on biopharma companies to boost product
titres.
58
Table 1.6: FDA approved therapeutic antibodies and Fc fusion proteins
Generic
name
Format Target Disease MOI Class Effector
function
Campath Humanized IgG1κ CD52 CLL Induction of ADCC I High
Erbitux
Chimeric
(murine/human)
IgG1κ
EGFR
Metastatic
colorectal
cancer/head
and neck
cancer
Inhibition of EGFR
signalling, ADCC
I
High
Arzerra Human IgG1κ CD20 CLL Induction of CDC
and ADCC
I High
Rituxan
Chimeric
(murine/human)
IgG1κ
CD20
Non-
hodgkin’s
lymphoma,
RA and CLL
Induction of
apoptosis, CDC and
ADCC
I
High
Herceptin Humanized IgG1κ HER2 HER2+
breast cancer
Inhibition of HER2
signalling, ADCC
I High
Humira
Human IgG1κ
TNFα
RA,
JIA,PA,CD,
AS and
plaque
psoriasis
Neutralisation of
TNFα activity
II
Moderate
Simulect
Chimeric
(murine/human)
IgG1κ
CD25
Acute
transplant
rejection
Inhibition of IL-2-
mediated activation
of lymphocytes
II
Moderate
Raptiva
Humanized IgG1κ
CD11a
Plaque
psoriasis
Inhibition of CD11a-
associated leukocyte
adhesion
II
Moderate
Simponi Human IgG1κ TNFα RA, PA and
AS
Neutralisation of
TNFα activity
II Moderate
Remicade
Chimeric
(murine/human)
IgG1α
TNFα
CD, RA, PA,
ulcerative
colitis, AS
and plaque
psoriasis
Neutralisation of
TNFα activity
II
Moderate
Enbrel
TNFR2 ECD-Fc
(IgG1) fusion
TNFα
RA, JIA,
PA, AS and
plaque
psoriasis
Neutralisation of
TNFα activity
II
Low
Vectibix
Humanized IgG2κ
EGFR
Metastatic
colorectal
carcinoma
Inhibition of EGFR
signalling: Induction
of apoptosis
II
Low
Avastin
Humanized IgG1κ
VEGF
Metastatic
colorectal
cancer
Neutralisation of
VEGF activity
III
Low
Xolair Humanized IgG1κ IgE Allergic
Asthma
Inhibit binding of IgE
to FcεRI
III Low
Table 1.6: Of the list of antibody therapeutics listed, Class refers to the specific mode of action (MOA)
specified further in figure 1.9.
59
1.4.1 Antibody structure and effector functions
Antibodies form an intricate part of the adaptive immune system and are responsible for
the recognition of potentially any antigen (epitope) through the three variable regions
within both the light and heavy chain that form the complementarity-determining region
(CDR) (Sela-Culang, Kunik & Ofran 2013) of the antigen-binding fragment (Fab). The
basic structure of a Mab consists of four polypeptide chains: two identical light chains
and two identical heavy chains (EDELMAN, BENACERRAF 1962). Figure 1.8 shows
the association that the four polypeptides demonstrate though covalent and non-covalent
interactions to form the three independent mature antibody structures. The two Fabs are
composed of the light chain and the N-terminal domain of the heavy chain harbouring the
variable regions which determines hyper-variability in antigen specificity along with two
constant domains (CH1 and CL). Additionally, two domains forming the heavy chains
are composed of two constant domains (CH2 and CH3) which constitute the crystallisable
fragment (Fc) and contain moieties that interact with ligands on immune cells that elicit
effector functions.
60
Figure 1.8: Structure of the immunoglobulin G (IgG) molecule
Figure 1.8: Various domains the makes up a mature IgG molecule. The two heavy chains and two
light chains are represented as the orange and blue structures, respectively, linked non-covalently by
the flexible hinge region. Antigen specificity is dictated by the two antigen-binding fragments (Fab)
made up of covalently linked light and heavy chain. Glycosylated sites are located on the conserved
Asparagine 297 (Asn297) on each heavy chain CH2 domain.
Fc receptors (FcRs) are glycoproteins belonging to the immunoglobulin superfamily and
are composed of the classes FcαR, FcγR and FcεR of which FcγR are mostly expressed
on immune cells and are fundamental for therapeutic outcome. FcγRs are divided into
three subtypes – FcγRI (CD64), FcγRII (CD32) and FcγRIII (CD16) categorised based
on the their affinity for IgG and the downstream signalling pathways that they trigger
(Nimmerjahn, Ravetch 2008). As mentioned earlier, antibodies can be exploited to elicit
a series of therapeutic outcomes such as neutralisation of a soluble signalling molecule
such as TNFα in Rheumatoid arthritis (Class III), opsonisation and clearance through
ADCC/CDC via targeting of cell-bound antigens (Class I) and receptor blocking
interfering with “normal” aberrant signalling (Class II), Figure 1.9.
61
Figure 1.9: Mechanism of action of therapeutic antibodies.
Figure 1.9: Mechanisms of action (MOA) of therapeutic antibodies including targeting and
subsequent clearance of cell/molecule via antigen binding and ADCC/CDC activation (Class I),
Interference with aberrant signalling through receptor binding and blocking (Class II) and
Clearance and deactivation of soluble antigens (Class III). Figure was sourced from (Jiang et al.
2011).
Additionally, better understanding of antibody structure and function have allowed for
the development of Fc-fusion proteins, such as Etanercept (Enbrel), which is a Tumour
Necrosis factor receptor 2 (TNFR2) fused to the constant Fc-region of an antibody which
prolongs its half-life in serum via its interaction with the neonatal Fc receptor and appears
to act independently of glycoforms (Roopenian, Akilesh 2007). The predominant subclass
of antibody licensed for therapeutic rMab use is the IgG1 isotype (Bruggemann et al.
1987) largely due to its strong association with the FcγRIIIa, a ligand responsible for
ADCC activation through natural killer (NK) cells (Bruhns et al. 2009). This affinity is
62
shared by the IgG3 isotype but not with the IgG2 and 4 subclasses. For example, the
primary pathway of Ritumximab in the treatment of CD20+ B lymphocytes in non-
Hodgkin’s lymphoma appears to be mediated through the activation of ADCC via
association with NK cells expressing FcγRIIIa receptors (Congy-Jolivet et al. 2007). It is
interesting that a clinically approved EGF-R antibody, panitumumab of the IgG2 isotype,
interacts less efficiently with human FcγRs and does not activate the compliment cascade
(Dechant et al. 2007). It is from these advancements in our understanding of Mab/immune
system relationship that derivation away from the predominant IgG1 isotype is a growing
field. Long term exposure to such antibodies in the case of chronic inflammatory diseases
can, by nature, activate immune cells to release inflammatory cytokines thus contributing
to the symptoms being targeted. Development of Mabs or fusion proteins specific for
TNFα but with a low affinity for FcγRIIIa receptors could function to sequester and block
TNFα signalling without, in this instance, the unwanted immune signalling.
1.4.2 Antibody glycoforms
Despite the CHO cell being the favoured production vessel of recombinant protein
therapeutics due to similarities in their post-translational modifications, the recent release
of the CHO genome sequence has revealed a number of discrepancies in the genomic and
transcriptomic profile of genes associated with glycan synthesis and degradation (Xu et
al. 2011). Although only constituting 2-3% of the total antibody mass, glycosylation of
the IgG-Fc region has been demonstrated to be essential to the activation of the
downstream biological effector functions. This glycosylation is through covalent
attachment of an oligosaccharide at the conserved Asparagine 297, Asn297 (Figure 1.8).
A diantennary type oligosaccharide comprised of a core heptasaccharide and variable
addition of fucose, galactose, bisecting N-acetylglucosamine and sialic acid contributes
to the glycoforms of Fc modified Mabs (Fig. 1.10). With the variation in the addition of
sugars to the core heptasaccharide, a total of 32 unique oligosaccharides can be generated
from each partner of the heavy chain. Taking into account the duplicity of heavy chains
in a single Mab there is the potential of ~500 glycoforms to be generated of mature
antibody (Jefferis 2009a).
63
Figure 1.10: Diantennary-complex oligosaccharide composition of the IgG-Fc
Figure 1.10: A schematic representation of the Diantennary-complex oligosaccharide that is
covalently attached to the conserved Asparagine 297 (Asn297) on both CH2 domains of the heavy
chains. The oligosaccharide is composed of a core heptasaccharide (sugars in blue) and outer arm
sugars (in red) variably attached, Fucose (Fuc), Galactose (Gal), Bisecting N-acetylglucosamine
(GlcNAc), sialic acid and N-acetylneuraminic acid (Neu5Ac). This figure was sourced from (Jefferis
2009a).
Nomenclature for naming the various glycoforms can use G0, G1 and G2 for core
oligosaccharides bearing zero, one or two galactose residues, respectively. In the event of
an addition of fucose (Fuc), the notation G0F, G1F and G2F is used. Finally, with the
addition of a bisecting N-acetylglucosamine (GlcNAc) a B is added to the naming system
previously described, such as G0B or G0BF (Fig. 1.11). Less than 10% of
oligosaccharides are sialylated so have not been included in the notation description
(Mizuochi et al. 1982).
64
Figure 1.11: Diantennary oligosaccharide structure and nomenclature
Figure 1.11: The naming system of the variable oligosaccharides that can be generated on each
isolated heavy chain and their corresponding glycan structure. Above shows the 16 variants of each
core heptasaccharide without including the addition of a sialic acid group. Taking into account the
16 variations on both heavy chains shown above, the total number of unique glycoforms is 128 (16 x
16) which is divided by two because of the symmetry of the molecule. This figure was sourced and
modified from (Jefferis 2009a).
It has been observed that the oligosaccharide present on glycoproteins contribute to their
stability and solubility (Krapp et al. 2003, Mimura et al. 2000) so it is not surprising that
a-glycosylated or de-glycosylated IgG forms result in a compromised or ablated effector
mechanism mediated through FcγRI, FcγRII and FcγRIII.(Sinclair, Elliott 2005). CHO
cells predominantly produce the top eight glycoforms (a-h) in figure 1.11; however, the
proportion of galactosylated and non-fucosylated IgG-Fc is relatively low by comparison.
CHO cells are unable to add the bisecting N-acetylglucosamine as they do not express the
GnTIII gene but do possess a genomic homolog (Xu et al. 2011). In the absence of these
A - G0
C – G1
E – G0F
G – G1F
I – G0B
K – G1B
M – G0FB
O – G1FB
B – G1
D – G2
F – G1F
H – G2F
J – G1B
L – G2B
N – G1FB
P – G2FB
65
glycan modifications, CHO cells possess the ability to add other sugars that humans do
not retain thus contributing to immunogenicity such as the addition of galactose in an α(1-
3) linkage. These observations have led to the exploration and development of CHO cells
engineered to produce specific glycoforms that mediate enhanced effector functions.
1.4.3 Glycan-manipulation
Present CHO cell production vessels produce rMabs with over 90% fucosylated forms of
IgG-Fc. Engineered CHO cells to express the GnTIII gene to add the bisecting N-
acetylglucosamine residue on the IgG-Fc bound core heptasaccharide demonstrated a 2-
fold increase in ADCC for a rituximab-like product (Davies et al. 2001). Further
investigation revealed that this enhanced effector ADCC was mediated through the
inhibition of the subsequent addition of Fucose to the primary N-acetylglucosamine
residue through the prior addition of the bisecting GlcNAc in the golgi apparatus (Ferrara
et al. 2006). Targeting fucosyltation of the primary GlcNAc is now emerging as a viable
approach for cellular engineering strategies, a process that has been extended to
encompass the therapeutic antibody Herceptin with a great deal of success in the
laboratory (Suzuki et al. 2007b).
A multitude of strategies have been employed in the engineering of CHO cells that
produced non-fucosylated antibodies such as short-hairpin RNA technology or genetic
knockouts of the gene responsible for mediating fucosylation, α1,6-fucosylatransferase
(FUT8) generating Mabs with over 100-fold increased potency in ADCC (Mori et al.
2004, Malphettes et al. 2010, Yamane-Ohnuki et al. 2004). Alternative strategies have
attempted to hit the fucosylation process from multiple angles including simultaneous
knockdown of the enzyme responsible for generating the prerequisite, GDP-fucose, for
fucosylation GDP-mannose 4,6-dehydratase (GMD), including the GDP-fucose
transporter and FUT8 with great success (Kanda et al. 2007, Imai-Nishiya et al. 2007).
An interesting observation was that in the case of GMD-only knockouts, fucosylation was
revived in the presence of L-Fucose supplementation through a salvage pathway that
bypassed the glucose-derived GMD-mediated generation of GDM-fucose (Fig. 1.12).
66
Figure 1.12: GDP-fucose salvage pathway for FUT8-mediated fucosylation
Figure 1.12: Antibody fucosylation can be targeted with the addition of compensatory pathways that
exist through the salvage pathway that generates GDP-fucose through the supplementation of L-
fucose which compensates for GMD knockout. The schematic diagram above was reported in (Kanda
et al. 2007).
The increased affinity of non-fucosylated IgG-Fc for FcγRIIIa may be due to the
reduction or absence of steric inhibition at the receptor-ligand interface. The structure of
the Fc portion of IgG is flexible and takes the form of a horseshoe-like conformation that
forms a binding cleft for FcγR interaction. The oligosaccharides covalently linked to
Asn297 are also flexible in nature and move to accommodate the IgG-Fc-FcγR
association. Fucosylation has been proposed to reduce the flexibility of the Diantennary
glycan thus contributing to this stoichiometric interference.
Despite the advancements in cell line development for the production of tailor designed
antibody glycoforms eliciting enhanced effector functions, implementing such
technologies to the manufacture of currently available recombinant Mabs such as
67
Herceptin and Rituxan would be a futile endeavour. The reason for this is that, although
the process would ultimately generate a more potent product, the resources and time
required to develop the engineered production CHO cell lines would be extensive.
Because the glyco-profile of the product being different to its predecessor despite antigen
specific remaining unchanged, it would be classified as a “New” product thus requiring
clinical trial testing, at high a cost, before FDA approval. All is not lost however, with the
patents expiring on many of these blockbuster therapeutics and over 350 novel and
competing therapeutic Mabs in early development, CHO cell production factories
genetically engineered to produce specific glycoforms can be implemented from the
ground up, playing a pivotal part in the generation of “Bio-similar” or hopefully “Bio-
Betters” (Reichert 2013).
68
Aims of thesis
1 Functional screen of a panel of miRNAs identified in a multi-omics profile of CHO
cells with varying growth rates
To:
Determine whether the integration of multiple datasets (miRNA, mRNA and
protein) can allow for the identification of high priority miRNA targets.
Evaluate the role miRNAs play in the regulation of CHO cell growth and other
bioprocess relevant phenotypes (Apoptosis and productivity).
Explore the feasibility of screening combinations of miRNAs.
Prioritize a short list of mRNA candidates desirable for stable CHO cell
development.
2 miR-34a as a potential candidate for CHO cell engineering
To:
Explore the stable depletion of miR-34 using miRNA sponge technology on the
CHO cell phenotype.
Determine if miR-34 depletion and miR-34a overexpression alters secreted
antibody fucosylation profiles.
3 miR-23b* as a potential candidate for CHO cell engineering
To:
Explore the influence of stable depletion of miR-23b* using miR-sponge
technology on the CHO cell phenotype.
4 miR-23 depletion as an attractive tool for industrial application
To:
Identify the role miR-23 depletion alone plays in CHO-SEAP cells following the
observation of enhanced growth/productivity under stable cluster inhibition.
Determine if miR-23 depletion enhances CHO cell mitochondrial function.
Determine the mechanism behind the observation of elevated SEAP activity.
Attempt to identify potential targets of miR-23 through label-free mass
spectrometry.
Carry out label-free mass spectrometry on isolated mitochondria.
69
Chapter 2
Materials & Methods
70
2.1 General Techniques of Cell Culture
2.1.1 Ultrapure water
Ultrapure water was used in the preparation of all media and solution when specified. For
ultra-purification, the water was initially pre-treated which involved activated carbon,
pre-filtration and anti-scaling. This water was then purified by a reverse osmosis system
(Elga USF Maxima Water Purification System) to a standard of 12 – 18 MΩ/cm
resistance.
2.1.2 Glassware
All glassware and lids were soaked in a 2% (v/v) solution of RBS-25 (VWR International)
for at least 1 h (hr). This is a deproteinising agent, which removes proteineous material
from the bottles. Glassware was scrubbed and rinsed several times in tap water. The
bottles were then washed by machine using Neodisher detergent, an organic, phosphate-
based acid detergent. Bottles were then rinsed twice with distilled water, once with
ultrapure water and sterilised by autoclaving (121˚C for 20 min under a pressure of 1bar).
The spinner vessels were additionally treated with 1M NaOH solution over night to
further ensure the elimination of cell debris from the inner surface and were then washed
twice with ultrapure water. Spinners were sealed in autoclave bags (Stericlin, 0775) and
autoclaved. Thermo-labile solutions (i.e. 10% DMSO, serum) were filtered through a 0.22
µm sterile filter (Millipore, millex-gv, SLGV-025BS).
2.1.3 Cell culture cabinet
All cell culture work carried out in a class II laminar air-flow (LF) cabinet (Holten).
Before use, the laminar was turned on and the air flow was allowed to acclimatize for 10
min. Both before and after use, the laminar flow cabinet was cleaned with 70% industrial
methylated spirits (IMS). Any items brought into the LF cabinet were swabbed down with
IMS and regular cleaning of gloved hands with IMS was carried out to insure the
prevention of contamination. Only a single cell line/clone was used in the LF cabinet at
any one time. If a second cell line/clone was being used a period of 15 min clearance was
adhered to in order to prevent/minimize cross contamination between cell lines/clones.
71
Weekly cleaning of the LF cabinet included Virkon, a detergent solution (Virkon, Antec
International; TEGO, TH. Goldschmidt Ltd.), followed by water and IMS.
2.1.4 Incubators
Adherent cells were maintained at 37 ̊ C, in an atmosphere of 5% CO2 and 80% humidity.
Cells in suspension were also maintained in these conditions in an ISF1-X (Climo-
Shaker) Kuhner incubator with a speed of 130-170 rpm. Weekly cleaning of the
incubators was adhered to and followed the same protocol as described for the cell culture
cabinet.
2.2 Subculture of cell lines
Cells were maintained as described above (see 2.1.4: Incubators). Cells were sub-
cultured every 3-4 days in order to maintain them in mid-exponential phase of their
growth cycle. In the event of temperature shift assays, cells were transferred from ambient
37 ˚C growth to a 31˚C incubator. All cell lines used and their required culture conditions
are tabulated below (Table 2.1).
Table 2.1: Cell lines, Culture conditions, Media types and Selection reagents
Cell lines and their associated growth conditions
Cell line
I.D.
Cell Type Culture
Media
Culture
Format
Antibiotic
Selecting
Agent
Antibiotic
Concentration
(nM/mL)
CHO-K1-
SEAP
rCHO CHO-S-
SFMII
Suspension Geneticin
(G418)
1000 µg
CHO-K1 CHO ATCC
+5%FCS
Monolayer None None
CHO-
1.14
IgG-
secreting-
rCHO
R5 Suspension Methorexate
(MTX)
450 nM
CHO-
5.18F
IgG-
secreting-
rCHO
AS1 Monolayer MTX 450 nM
CHO-
11.1S
IgG-
secreting-
rCHO
AS1 Monolayer MTX 450 nM
CHO-
5.22F
IgG-
secreting-
rCHO
AS1 Monolayer MTX 450 nM
72
2.2.1 Anchorage-dependent cells (Monolayer)
Spent media was removed from the flask by pipette and discarded into a sterile waste
bottle. The flask was then rinsed with 2-5 mL of autoclaved/sterile PBS in order to remove
any residual media. Depending on the size of the culture vessel, 2-5mL of trypsin was
added to the flask covering the cells and incubated at 37 ˚C for approximately 2-5 min
until all of the cells have become detached. The level of cell detachment was monitored
by microscopic observation. A 5mL volume of serum-containing media was added to the
flask once all the cells were detached in order to deactivate the trypsin. The cell
suspension was placed into a 30 mL sterile universal container and centrifuged at 1000
rpm for 5 min. The supernatant was discarded from the universal and the pellet was
resuspended in fresh complete media. A small aliquot was removed and cell count was
performed. Another aliquot of this cell suspension was then removed and used to seed a
fresh culture vessel at the required density. Cells were passaged once 90-100%
confluency was reached otherwise spent media was removed and fresh media was added
to the culture flask.
2.2.2 Suspension cells
Stock cells were maintained in a spinner vessel in a 30-50 mL volume at 37 ̊ C in a Kuhner
Climo-shaker at 170 rpm. A small aliquot was sampled for counting by trypan blue
exclusion and the required seeding density of 2 x 105 cells/mL was calculated. If the
required volume of cells from the working stock exceeds 10% of the culture volume then
the amount of cells needed is centrifuged at 1000 rpm for 5 min. Conditioned media is
poured off into a sterile waste container and the cell pellet is resuspended in 30-50 mL of
media. Some cells lines require the fresh stock flask to be supplemented with 10%
conditioned media.
2.2.3 Cell counting and viability determination
2.2.3.1 Trypan blue exclusion method
Prior to cell counting, an aliquot of sample was extracted from either the subculture of
adherent cells or a direct sampling from suspension cells as described in (Subculture
73
2.2.2). An aliquot of cell suspension was added to Trypan Blue (GIBCO, 525) at a ratio
of 1:1. Prior to the addition of Trypan Blue a dilution can be performed by adding volume
of media to the sample. The Trypan Blue/Cell sample mixture was incubated at room
temperature for 2-3 min. A small aliquot (10 µL) was then applied to the chamber of a
glass cover-slip enclosed haemocytometer. For each of the four corner grids of the
haemocytometer, cells in the 16 squares were counted under a light microscope (Fig. 2.1).
Figure 2.1: Haemocytometer Grid Layout
The average of the four grids was multiplied by a factor of 104 (Volume of the grid) and
the relevant dilution factor to determine the average cell count per mL in the original
suspension. Viable cells exclude the Trypan Blue dye as their membranes remain intact
and therefore remain unstained while non-viable cells stain blue. The percentage viability
of the culture is calculated by counting non-viable cells and viable cells in the sample and
dividing the total cell count (viable and non-viable) by the viable cell count.
2.2.3.2 Guava ViaCount® assay
The Guava ViaCount assay is an efficient alternative to Trypan Blue Exclusion for the
high throughput determination of cell numbers and viability. The active Guava
74
ViaCount® Reagent (Guava Technologies, 4000-0041) discriminates between viable and
non-viable cells based on the permeability of two DNA-binding dyes. One dye is
membrane permeable and stains all nucleated cells producing a fluorescent signal that is
detected by photomultiplier tube 2 (PM2). The remaining dye penetrates the membrane
of non-viable cells of which membrane integrity has been lost producing a detectable
fluorescent signal on photomultiplier tube 1 (PM1). Single positive events of viable cells
are detected and recorded when the fluorescent signal is accompanied by a signal of
forward light scatter (FSC) which is of an appropriate intensity to that produced by a cell
above a particular size. An Event with an accompanying low FSC is counted as cellular
debris and not included within the live nucleated cell population. Non-viable cells will
emit a fluorescent signal form the dead cell stain. The read out format for the Guava
ViaCount assay is a 96-well plate and no less than 100 µL should be used per well. It is
essential that the cell count present is no less than 10 cells/µL and no more than 500
cells/µl which could result in inaccurate reads. Cells were diluted in serum free media as
well as taking into account the addition of 100 µL of Guava ViaCount® Reagent. Plate
set-up and inclusion of dilution factors is described in section (see 2.2.3.2.1). Guava
ViaCount® Reagent was allowed to adjust to room temperature. 100 µL of pre-diluted
sample was added to 100 µL of Guava ViaCount® Reagent and allowed to incubate at
room temperature for at least 10 min before using the Guava flow cytometer.
2.2.3.2.1 WorkEditTM Software
The WorkEditTM software programme allows the acquisition command set-up to be
performed for samples in a 96 well Round Bottom microplate (Corning Inc. COSTAR®,
10308005) in the sample tray. Figure 2.2 shows the screen layout of the WorkEditTM
command template. The right hand side of the screen shows the grid template of the 96
well microplate and allows for the selection of sample wells highlighted in green. The left
hand side of the screen allows for the selection of the assay type to be performed in this
case being Guava ViaCount. Acquisition events and acquisition order allows the selection
of the number of viable cell events per sample and the order in which the microplate will
be read. Most importantly, entering the dilution factor is necessary for the analysis of each
samples cell numbers accurately.
75
Figure 2.2: WorkEditTM command screen layout
Figure 2.2: WorkEditTM command screen layout: The right hand side shows the sample well layout
and allows selection of the well for analysis (highlighted in green). Above it the layout for 10
microcentrifuge sample tubes along with the 6 well locations required for cleaning. The left hand side
of the screen allows the selection of assay type, acquisition event number, acquisition order, auto
saving data files, mid run washing steps, and sample dilution factor.
2.2.3.2.2 EasyFit Analysis
EastFit analysis allows the software to calculate the results based on its own methodology.
This type of analysis groups all events into three individual population categories – viable
cells, dead/apoptotic cells and cellular debris. EasyFit uses its own grouping method that
clusters data into populations discriminating between the three population types. A
manual method can be executed alongside this where a gate can be imposed upon the
“healthy” population by altering the FSC threshold thereby completely discriminating
between cellular debris and viable cells.
76
2.2.3.3 Cedex XS Analyser
The Cedex XS Analyser (Roche Innovatis, Bielefeld, Germany) is a semi-automated
image-based cell analyser based on the Trypan Blue Exclusion method providing
information about cell concentration and viability. The Cedex XS Smart Slide has eight
chambers for testing samples. Four sample chambers (e.g. Chambers A1-A4) can be
measured without turning the Cedex XS Smart Slide. Cell samples for measurement are
pre-diluted (below 1x107 cells/mL) using and stained using the appropriate amount of
0.2% Trypan Blue while accounting the addition of Trypan Blue into the dilution factor.
This solution is mixed thoroughly by pipetting up and down. Pipette 10 µL of sample into
each chamber of the Cedex Smart Slide for measurement (Figure 2.3 A). The slider
carrier is pulled out and the Cedex Smart Slide is placed in the slide carrier with the
chambers containing the samples going in first. The slide carrier is pushed gently in until
a soft click is felt followed by the illumination of a blue light indicating the first chamber
is in position, with an additional light for chambers 2, 3 and 4. By clicking on MEASURE
from the Cedex XS control centre the measurement dialogue box is accessed where
sample information is entered such as sample ID including replicates and dilution factor
before taking the final measurement.
Figure 2.3: Cedex Smart Slide & Sample view
Figure 2.3: Cedex smart slide and Sample view: A) Sample loading of 10 µL of pre-diluted sample
into the chamber of the Cedex Smart Slide. B) Sample view accounting for viable cells circled in
green, non-viable cells circled in red and aggregates circled in blue.
A B
77
Once the sample parameters have been entered the START MEASUREMENT button is
clicked. The Cedex XS analyser accounts for information such as cell density by taking
the dilution factor into account, cellular viability and others such as cell size ranging from
2-180 µm and aggregates (Figure 2.3 B).
2.2.4 Cryopreservation of cells
Cells being preserved were frozen down in a suitable medium supplemented with 5%
DMSO (SIGMA-ALDRICH, D5879) and 5% FCS. Given the cytotoxic effects of DMSO
at warm temperatures a separate solution of suitable medium supplemented with 10%
DMSO/FCS was made up initially and kept on ice until required. This 10% DMSO/FCS
solution was filtered using a 0.22 µm bell filter (Gelman, 121-58) before storage on ice.
Cells for cryopreservation were harvested during their log phase of the growth cycle.
Generally, cells are frozen down at 1x106 cells/mL in 1mL of media (supplemented with
5% DMSO/FCS) in a cryovial (Greiner, 201151). After counting and centrifugation, cells
are resuspended in half the final volume of pre-warmed suitable media giving a cell
density at 2x106 cells/mL. Finally, this mixture is made up with 1 volume of the 10%
DMSO/FCS supplemented media yielding a final solution of 5% DMSO/FCS at 1x106
cells/mL. This suspension was then aliquoted in 1mL volumes to pre-labelled cryovials
and immediately placed at -20˚C. After 1-2h, cells are transferred to -80˚C for 6-12 h and
then transferred to the liquid nitrogen tank for long term storage at -196˚C.
2.2.5 Thawing of cryopreserved cells
Prior to thawing of cells under liquid nitrogen storage, a 10 mL volume of fresh medium
was pre-warmed in a sterile universal. The cryopreserved cells were removed from the
liquid nitrogen storage, immediately placed on ice and slightly thawed at 37 ˚C. Cells
were transferred from the cryo-vial to the aliquoted medium in the sterile universal using
a pastet. Cells were centrifuged at 1000 rpm for 5 min in order to remove the toxic DMSO.
The supernatant was decanted and the pellet was resuspended in fresh pre-warmed media.
Antibiotic selection agent is omitted in the first passage. Thawed cells were then added
to the appropriate culture vessel, i.e. spinner or T-flask, with the appropriate volume of
culture medium and allowed to grow at 37 ˚C. In the case of adherent cells, the culture
78
media was removed and cells were re-fed the next day in order to remove and non-viable
cells and allow the successful proliferation of attached cells.
2.2.6 Apoptosis analysis
Annexin-V is a calcium-dependent phospholipid binding protein that binds with high
affinity to phosphotidylserine (PS), which is normally located on the inner face of the cell
membrane. Annexin V-PE is used to detect PS on the external face of the membrane and
is a sign of early apoptosis. 7-AAD is also membrane impermeable and is excluded from
live, healthy cells and early apoptotic cells.
Non-apoptotic cells: Annexin-V (-) and 7-AAD (-)
Early Apoptosis: Annexin-V (+) and 7-AAD (-)
Late stage apoptosis/Dead: Annexin-V (+) and 7-AAD (+)
Debris: Annexin-V (-) and 7-AAD (+)
Cells were harvested and resuspended in 200 µL Guava Nexin reagent (Merck-Millipore)
at 48 h and 96 h after transfection. Cells were incubated for 20 min at room temperature
in the dark before analysis with a Guava EasyCyte96 (Millipore). It was ensured that at
least 1% of FCS was added to the cell sample before incubation in cases of serum-free
culture.
2.3 Other Cell Culture Techniques
2.3.1 Reconstitution/Rehydration of lyophilized miRNA
miRNAs were ordered from Applied Biosystems and NBS Biologicals and arrived in a
lyophilised condition. Reconstitution/Rehydration of miRNA was performed using
nuclease-free water (Ambion®, AM9932) in sterile conditions in the laminar flow hood.
miRNAs were resuspended so that their final stock concentration was at 50 µM. The
appropriate volume of nuclease-free water was added by pipette and mixed vigorously by
pipetting up and down and vortex. This mixture was allowed to sit at room temperature
for 10 min and subsequently centrifuged briefly before storage at -20˚C. Stocks were also
79
subsequently made as low as 5 µM in the event of transfecting small concentrations of
miRNA, allowing for high volumes being used to avoid pipetting error.
2.3.2 siRNA/miRNA transfection
2.3.2.1 siSPORTTM NeoFXTM Transfecting Reagent
Transient transfection of CHO cells was carried out using the AMBION® transfecting
reagent siSPORTTM NeoFXTM (AMBION®, AM4510) in accordance with the
manufacturer’s specifications. The following protocol describes the experimental setup
for a single transfection event at a siRNA/miRNA concentration of 50 nM including a
10% excess. Each siRNA/miRNA/NeoFX complex volume of 200µl was added to a 2mL
sample of cells at 1x105 cells/mL. All sample sets were carried out in triplicate with
appropriate experimental controls. Using serum free media (SFM) such as OPTI-MEM®
(Life Technologies, 31985-062) or CHO-S-SFM II (GIBCO®, 12052-098), 2.2 µL of a
50µM miRNA/siRNA was added to 110 µL of SFM previously warmed to 37 ˚C and
incubated at room temperature for 10 min. The siSPORTTM NeoFXTM transfecting
reagent was allowed to sit and adjust to room temperature before the addition of 2.2 µL
to 110 µL of pre-warmed SFM in a separate microcentrifuge tube. Formation of
siRNA/miRNA complexes was achieved by adding 112.2 µL of siSPORTTM NeoFXTM
transfecting Reagent/SFM media mixture to the incubated siRNA/miRNA/SFM sample
and allowed to complex by incubation for 10 min at room temperature. In order to
inoculate the 2 mL (1x105 cells/mL) culture with a 50nM final concentration of
siRNA/miRNA 204 µL of the final complex was added with gentle swirling of the culti-
flask disposable bioreactor (Sartorius Stedium Biotech, DF-050MB-SSH). Cells were
incubated at 37 ˚C in a Kuhner Climo-shaker incubator (Kuhner, Climo-Shaker ISF1-X)
at 170 rpm. Replacement of the culture media is not necessary as the addition of
siSPORTTM NeoFXTM transfecting Reagent does not induce cellular stress that can lead
to cytotoxicity.
80
2.3.2.2 INTERFERinTM transfection reagent
Transfection of CHO cells was carried out using the PolyPlus-transfectionTM reagent
INTERFERinTM (PolyPlus-TransfectionTM, PPLU-409-01) as per manufacture’s
specifications unless otherwise indicated. miRNAs were transfected at a range of 100 nM-
20 nM final concentration per 2 mL volume at 1 x 105 cells/mL. Below describes the
master mix preparation for one single transfection. Using SFM such as OPTI-MEM®
(Life Technologies, 31985-062) or CHO-S-SFM II (GIBCO®, 12052-098) 110 µL of pre-
warmed SFM media at 37 ˚C was aliquoted into a sterile cryovial or Eppendorf. All
reagents such as INTERFERinTM and miRNAs were kept on ice until required, however
INTERFERinTM was allowed to adjust to room temperature before use. As described in
section 2.3.1.1, 2.2 µL of the appropriate stock miRNA mimic or inhibitor was added to
the aliquot of 110 µL SFM and mixed by pipetting up and down gently. From
optimization results shown in section 3.1.2.4, 1.1 µL of INTERFERinTM was added to
the SFM containing miRNA mixture and allowed to incubate at room temperature for ~
15 min. After this incubation period, 100 µL of complexed miRNA was added directly to
each 2 mL volume of cells and swirled briefly to mix. Cells were incubated at 37 ˚C in a
Kuhner Climo-shaker incubator (Kuhner, Climo-Shaker ISF1-X) at 170 rpm.
Replacement of the culture media is also not necessary as the addition of INTERFERinTM
transfecting reagent does not induce cellular stress that can lead to cytotoxicity.
2.3.3 Transfection of plasmid DNA
Plasmid transfection required the use of a separate set of transfection reagents as plasmids
are required to reach the nuclear membrane thus having to bypass two membranes.
2.3.3.1 Lipofectamine 2000
Cells were transfected at 5 x 105 attached 24 hs prior to transfection or at 0.5-1 x 106
cells/mL in suspension culture. DNA to LipofectamineTM (Invitrogen, Cat. No. 11668-
019) was 1:2, respectively. DNA was diluted in 50 µL SFM and mixed gently.
Transfection reagent was diluted in a separate 50 µL, mixed and incubated for 5 min at
room temperature. After 5 min, both mixtures were combined, mixed and incubated at
81
RT for 25 min. After 25 min, the 100 µL mixture was added drop-wise to the cells and
gently rocked to mix. This transfection reagent requires a media change after 4-6 hs of
incubation.
2.3.3.2 TransIT®-2020 Transfection Reagent
Cells were transfected at 5 x 105 attached 24 hs prior to transfection or at 0.5-1 x 106
cells/mL in suspension culture. Transfection complexes were made up in SFM such as
CHO-S-SFMII or Optim-MEM. For a single transfection, 250 µL of SFM was pre-heated
to 37°C in a 1.7 mL Eppendorf. 1 µL of TransIT-2020 (Mirus, Cat. No. MIR-5400) was
used for every 1 µg of plasmid DNA transfected. DNA was added first using a pipette
and mixed followed by the transfection reagent. Contents were mixed and allowed to
incubate at room temperature for 20 min. After this time, DNA/transfection reagent
complex was added drop-wise to the cells. This transfection reagent does not require a
media change.
2.3.4 Limited-dilution/FACS-based cloning
Clonal isolation was carried out using a FACS-Aria based on mid-to-high GFP intensity
or limited dilution single cell cloning. The culture process after isolation was the same
for both processes. In the case of limited-dilution cloning, a stock solution is made up
containing 10 cells/mL so that when 100 µL is placed into a well, the well contains a
single clone. Clones were sorted into 2 separate 96 well plates containing CHO-S-SFM
II media supplemented with ~30% conditioned media (media previously cultured CHO
cells and harvested containing secreted growth factors) and 5% serum. Conditioned media
was harvested 2-3 days into culture with cells removed by centrifugation and filtered
through a 0.2 µm filter. Single clones were allowed to attach and grow at 37 ˚C for 10-14
days. After this time, cells were assessed for growth and clonality visually using a bench-
top microscope. Clonal panels were subsequently adapted to suspension culture in a 24-
well suspension plate.
82
2.4 Molecular Biology Techniques
2.4.1 DNase Treatment of RNA
DNase treatment of RNA samples is useful for the removal of contaminating genomic
DNA and for the removal of plasmid DNA from transfection reactions that would have,
in its high abundance, carry over during the RNA isolation protocol.
DNase treatment was carried out using the DNase I (RNase-free) kit (M0303S, from New
England Biolabs®) as per manufacturer’s specifications. Each DNase treatment reaction
was made up to a reaction volume of 20 µL. 2ug of RNA was added to a 0.5 mL PCR
tube as well as 2 µL of 10X DNase Buffer (resulting in a 1X DNase Buffer when made
up to 20 µL with RNase-free water). 2 Units of DNase I (1 µL) was added to the reaction
resulting in a final reaction volume of 20µl. The reaction mixture was mixed thoroughly
and incubated at 37 ˚C for 10-15 min. 1 µL of 0.5 M EDTA (to a final concentration of 5
mM) was added after this incubation step in order to deactivate the DNase I enzyme with
a subsequent heat inactivation step at 75˚C for 10 min.
2.4.2 High Capacity cDNA reverse transcription
Reverse transcription of RNA to cDNA was carried out using the High Capacity cDNA
reverse transcription kit (Applied Biosystems, 4368814) in accordance with the
manufacturer’s specifications. A maximum of 2 µg RNA was made up to 10 µL in a 0.5
mL PCR tube using Nuclease-free water. In a separate 0.5 PCR tube reaction vessel, the
reverse transcription master mix was prepared in accordance with the following table
(single reaction volumes):
Component Volume/Reaction (µL)
Kit with RNase
Inhibitor
Kit without RNase
Inhibitor
10X RT Buffer 2 2
25X dNTP Mix (100mM) 0.8 0.8
10X RT Random Primers 2 2
MultiscribeTM Reverse
Transcriptase
1 1
RNase Inhibitor 1 -
Nuclease-free H2O 3.2 4.2
Total per reaction 10.0 10.0
83
10 µL of the RT master mix was added to the 10 µL RNA sample and mixed using a
pipette. The temperature profile for the reverse transcription reaction carried out using a
bench top thermocycler G-Storm GS1 PCR machine following the below criteria:
Step 1 Step 2 Step 3 Step 4
Temperature
(˚C)
25 37 85 4
Time 10 min 120 min 5 min ∞
Reverse transcribed cDNA product is ready to use and is stored at -20˚C until further
required.
2.4.3 Polymerase Chain Reaction (PCR)
The DreamTaqTM PCR Master Mix kit (Fermentas Life Sciences, #K1071) was the kit
used to perform all PCR amplification reactions in accordance with the manufacturer’s
specification. DreamtaqTM PCR Master Mix was vortexed after thawing and placed on
ice until further required. Each PCR reaction was made up to a volume of 50 µL.
Following the component table below, a 25 µL reaction volume was made up in a thin
walled PCR tube using a maximum concentration range of 1-100 ng of genomic DNA or
cDNA and 10 pg of forward and reverse PCR primer. Once all components have been
added, 25 µL of the 2X DreamTaqTM PCR Master Mix was added to each sample creating
a 50 µL reaction volume.
Component Volume/Concentration
DreamTaqTM PCR Master Mix (2X) 25 µl
Forward Primer 0.1 – 1.0 µM
Reverse Primer 0.1 – 1.0 µM
Template DNA 1 – 100 ng
Nuclease-free water To 50 µL
Total Volume 50µL
Samples were gently vortexed and quickly spun down. The thermo-cycling conditions are
shown in the table below with variation lying within the specific melting temperature
(Tm) of each primer set and length of the Amplicon. A bench top PCR machine was used
to perform the polymerisation cycle.
84
Step Temperature, ˚C Time Number of cycles
Initial Denaturation 95 1-3 min 1
Denaturation 95 30s
25-40 Annealing Tm – 5 30s
Extension 72 1min/kb
Final Extension 72 5-15 min 1
HOLD 4 ∞ 1
After completion of the PCR reaction, samples were stored at -20˚C until further required.
Primer sequences for all PCR reactions are shown in table 2.2.
Table 2.2: Primer sequences for all genes explored in PCR
Gene ID Primer For-Rev (5’-3-) Amplicon
(bp)
Tm
(˚C)
GC% Ann
Temp
E2F3 For-
CAGAGGTGCTCAAAGTGCAG
175
59.4 55
56
Rev-
AGCTCGGTCACTTCTTTGGA
57.3 50
BCL2 For-
GGTGGTGGAGGAACTCTTCA
153
59.4 55
56
Rev-
CGGTTCAGGTACTCGGTCAT
59.4 55
CCND1 For-
ACACCAACCTCCTGAACGAC
214 59.4 55
56
Rev-
GACAGGAAACGGTCCAGGTA
59.4 55
MYCN For-
CAAGAGCGAGGTTTCTCCAC
202 59.4 55 56
Rev-
AGCCTTCTCGTTCTTCACCA
57.3 50
GLS For-
AAGTGCTTGAATGGGCTGTT
197 59.7 45 56
Rev-
AGCAGTGGGGTGGAGTGTAT
59.4 55
TFAM For-
ATCTGCTCCGGGAAGACTG
190 60.3 57 56
Rev-
GTCTGAGCGTCTCACATTGC
59.5 55
DICER1 For-
GGAGAACTTCAGCAGCCAAC
164 59.4 55
56
Rev-
TCAGGTGGTTCAGGGTTTTC
57.3 50
FUT8 For-
TTCATCCCAGGTCTGTAGGG
200 59.4 55
56
Rev-
TTCCAGCCACACCAATGATA
55.3 45
85
CCNE2 For-
CGCAGTAGCCGTTTACAAGC
177 59.4 55
Rev-
AGTCCACGTCTGCATTACCC
59.4 55
CDK4 For-
GGATTGCCCTCAGAAGATGA
179 57.3 50
56
Rev-
AGGGATTGGAAGGCAGAAAT
55.3 45
CDK6 For-
TCAAATGGCCCTTACCTCAG
160 57.3 50
56
Rev-
CACGTCTGAACTTCCACGAA
57.3 50
2.4.4 Fast SYBR® Green Real-time qPCR
Real-time qPCR was performed using the Applied BioSystemsTM Fast SYBR® Green
Master Mix (Applied BioSystemsTM, Cat. No. 4385612) in accordance with the
manufacturers’ specifications. Quantification of the target gene is based on the binding
affinity of SYBR® Green for double-stranded DNA. During PCR amplification by
AmpliTaq® Fast DNA Polymerase, the target sequence is amplified creating more double-
stranded Amplicon for SYBR® Green to bind to. As the PCR reaction proceeds, the
fluorescent intensity is proportional to the increase in double-stranded PCR product. 20
µL reaction mixtures were prepared in a MicroAmpTM Fast Optical 96-well Reaction plate
with Barcode (4346906). Both primers and cDNA were placed on ice to thaw with
subsequent vortexing and brief centrifugation. The Fast SYBR® Green Master Mix was
mixed thoroughly by swirling. Prior to plate setup, a master mix was prepared for all
samples and technical replicates including a separate master mix for the endogenous
control (β-actin or GAPDH). For a single reaction, 10 µL of Fast SYBR® Green Master
Mix (2X) was added to an Eppendorf tube along with 7 µL Nuclease-Free water and 1
µL of the appropriate forward and reverse primers (200 nM of each primer). Subsequent
to master mix preparation, 18 µL were added to each well of the MicroAmpTM reaction
plate. Using a micro pipette designed for small volumes (<10 µL), 2 µL of cDNA template
was added to each reaction well using a fresh filter tip for each aliquot. It is recommended
that a maximum of 20 ng of template cDNA be used for each reaction to achieve optimal
detection. The reaction plate was then sealed using MicroAmpTM Optical adhesive film
(4311971) to avoid loss of reaction volume during the PCR process. The sample plate
was read using a 7500 Real-time PCR system. Data analysis was based on the Relative
86
Quatitation (2-ΔΔCt) where the relative expression of the target gene is compared to the
expression of an internal standard, endogenous control.
2.4.5 Determination of DNA/RNA quantity and quality
Numerous methods are available for the determination of the quantity and quality of DNA
and RNA.
2.4.5.1 Bioanalyser/RNA
The Agilent 2100 Bio-analyser is a microfluidics-based platform used for the analysis of
proteins, DNA and RNA. The miniature chips are made from glass and contain a network
of interconnected channels and reservoirs. The RNA 6000 Nano LabChip kit enables
analysis of samples containing as little as 5 ng of total RNA. To each well, 1µl of each
sample is loaded along with a fluorescent dye (marker). An RNA ladder is loaded into
another sample well for size comparison. When all the wells are loaded, the chip is briefly
vortexed and loaded onto the bio-analyser machine. The machine is fully automated and
electrophoretically separates the samples by injecting the individual samples contained in
the wells into a separation chamber.
87
Figure 2.4: RNA 6000 Nano chip and RNA Integrity Number graphs
Figure 2.4: RNA 6000 Nano chip and RNA Integrity Number graphs: A) a picture of the front of the
RNA Nano chip. B) Upper panel represents the RIN values for high quality RNA while the lower
panel displays the RIN values for low quality RNA. Diagrams were sourced from
http://www.agilent.com/chem/labonachip.
The fluorescence was measured and presented in the form of peaks for 18S and 28S
ribosomal RNA peaks. In figure 2.4 B, the upper electropherogram and gel-like image
show the analysis of high quality total RNA resulting in a RNA Integrity Number (RIN)
of 9.2. The lower electropherogram shows the analysis of a partially degraded total RNA
sample. Many degradation products appear between the two ribosomal bands and below
the 18S band resulting in a RIN of 5.8. As RNA degrades, the 28S RNA peaks decrease
and the smaller fragments appear.
2.4.5.2 NanoDrop® Spectrophotometer
Purified RNA/DNA samples were quantified using the Nanodrop® ND-1000
spectrophotometer (NanoDrop Technologies). Before applying the RNA/DNA sample,
the pedestal was wiped down using a lint-free tissue dampened with UHP. A volume of
2 µL of UHP was then loaded onto the lower measurement pedestal. The upper arm was
then lowered down so as to be in contact with the RNA/DNA containing solution. The
A B
88
programme “Nucleic acid” and then “RNA-40” or “DNA-50” was selected on the
NanoDrop software main screen to read the samples at 260 nm. Upon selection of the
appropriate programme (RNA-40, DNA-50) the blank option was chosen, after a straight
line appeared on the screen with zero appearing in all measurement fields the “measure”
option was selected. Before and after loading each subsequent sample the pedestal was
repeatedly wiped with a lint-free wiped. The concentration of RNA/DNA was calculated
by software using the following formula:
OD260nm x Dilution factor x 40 = ng/µl RNA
A260/A280 and A260/A230 ratio for each sample was recorded. An A260/A280 ratio ~2 and an
A260/A230 of ~1.8-2.2 is indicative of a pure RNA/DNA sample with an absorbance
considerably above or below these values indicating the presence of contaminants such
as protein or phenol. However, RNA/DNA ratios of as low as 1.86 and 2.24 for A260/A280
and A260/A230 respectively have been used in subsequent experiments.
Sample limitation for accurate read outs of RNA/DNA is 2 ng/µL minimum and 3000
ng/µL maximum without dilution.
2.4.5.3 Gel Electrophoresis/RNA
Visual determination of RNA integrity can be identified using gel electrophoresis. RNA
samples were previously diluted in RNase free water with all agarose gel containers being
cleaned and wiped down with RNase Zap. A 1% agarose gel in 50 mL 1 x TAE with 5
µL ethidium bromide (EtBr) added for staining the RNA. A DNA ladder was added to
one well of the gel to estimate the location of the RNA bands. The gel was run for a
sufficient enough to time to ensure clear separation of the RNA bands. Two RNA bands
should be visualised, 28S and 18S, with 28S being roughly twice as bright as the 18S
band. Bands should appear discrete without any leading smear. No bands should appear
between the 28S and 18S bands indicating genomic contamination and/or RNA
degradation.
89
2.4.6 Total RNA isolation using TRI ReagentTM
Isolation of total RNA from cell lysates was performed using the TRI ReagentTM mRNA
isolation method (SIGMA-ALDRICH®, T9424) in accordance with the manufacturer’s
specifications. Total RNA was extracted from a maximum of 10 x 106 cells. Suspension
cells were harvested by determining the appropriate cell count per mL using Trypan Blue
Exclusion (GIBCO®, 15250-061) and spinning the appropriate volume of culture mixture
at 2,000 rpm for 5 min. Supernatant was removed from the cell pellet and cells were
subsequently lysed with the addition of 1 mL TRI ReagentTM, mixed well by pipetting
and transferred to a microcentrifuge tube. Cell lysates can be stored in TRI ReagentTM at
-70˚C for up to one month. Samples were allowed to stand for 5 min at room temperature
to allow complete dissociation of nucleoprotein complexes. Phase separation was
achieved with the addition of 200 µL of Chloroform (SIGMA-ALDRICH, C2432) to the
cell lysate in 1 mL of TRI ReagentTM. This mixture was mixed vigorously by vortexing
at high speed for 15 seconds and allowed to stand at room temperature for 10 min.
Separation of the three distinct phases (red organic phase containing protein, intermediate
phase containing DNA and an upper aqueous colourless phase containing RNA), was
achieved by centrifugation at 12,000 rpm for 15 min at 4˚C. The aqueous phase (upper-
clear) was transferred to a fresh microcentrifuge tube and 500 µL isopropanol (SIGMA-
ALDRICH, I9516) was added and mixed. This mixture was allowed to stand for 10 min
at room temperature and then subsequently centrifuged at 12,000 rpm for 10 min at 4˚C.
RNA will form a precipitate along the side/bottom of the tube. The supernatant was
removed and a wash step was performed using 1 mL of 75% ethanol. Sample was mixed
well by vortex and centrifuged at 7,500 rpm for 10 min at 4˚C. Decant 75% ethanol as
best as possible and air dry the RNA pellet by allowing sitting for 10-15 min. The RNA
pellet was not allowed to dry completely before re-suspension in 30-50 µL nuclease-free
water.
2.4.7 PCR based miRNA Taqman® Low Density Arrays (TLDA)
Taqman® Low Density Arrays (Applied Biosystems, 4334812) are 384-well micro fluidic
cards designed for the analysis of global miRNA expression patterns of up to 8 samples
in parallel across a defined set of miRNA targets (Fig 2.5). Taqman® Low Density Array
90
is a real-time PCR-based, highly sensitive and reproducible array that contains specific
probes for human miRNA signatures.
Figure 2.5: Schematic representation of a miRNA Taqman® Low Density Array
Given the large number of known human miRNAs, miRNA TLDA cards are split into
two cards (A + B) with probes for 380 unique miRNA on each card and 8 endogenous
controls. Total RNA was harvested from cells using the TRI Reagent protocol and cDNA
was prepared using the MultiplexTM RT Primer kit.
91
2.4.7.1 Reverse transcription
The Taqman® microRNA Reverse Transcription kit and the MegaplexTM RT Primers Pool
A and B (4399966 and 4399968) were used to synthesise single-stranded cDNA from
total RNA. The RT-master mix was prepared as described below:
RT Reaction Mix
Components
Volume for One Sample
(µL)
Volume for Ten Samples
(µL) with 12.5% excess
MegaplexTM RT Primers 0.80 9.00
100mM dNTPs with dTTP 0.20 2.25
MultiscribeTM Reverse
Transcriptase (50 U/µl)
1.5 16.88
10 X RT Buffer 0.80 9.00
MgCl2 (25mM) 0.90 10.12
RNase Inhibitor (20 U/µl) 0.10 1.12
Nuclease-Free water 0.20 2.25
Total 4.50 50.62
The above mixture was combined in a 1.5mL Eppendorf tube depending on the number
of sample reactions being reverse transcribed. The mixture was mix by inverting the tube
and briefly centrifuged. In a 96 well MicroAmp® Optical reaction plate (4314320), 4.5
µL of the RT mixture was added into each well as required. 3 µL of total RNA (100 ng)
was then added to each reaction well and the plate was sealed using MicroAmp® Clear
Adhesive Film. The plate was left to incubate on ice for 5 min. Multiplex reverse
transcription was performed using the thermocycler under the 9600 Emulation mode with
a final reaction volume of 7.5 µL in the HT7900 RT-PCR machine and programmed as
follows:
Step Type Time (min) Temperature (˚C)
HOLD 30 16
HOLD 30 42
HOLD 5 85
HOLD ∞ 4
cDNA can be stored at -15 to -20˚C for at least one week.
92
2.4.7.2 Pre-amplification
The Pre-amplification step is required for low concentrations of starting cDNA whereby
specific cDNA targets are amplified for further Taqman® MicroRNA array analysis. The
MegaplexTM PreAmp Primers (Pool A and B, 4399233 and 4399201) were thawed on ice,
mixed by inverting and briefly centrifuged. Master Mix was prepared in a 1.5 mL
Eppendorf tube in accordance with the following recipe:
PreAmp Reaction Mix
Components
Volume for One sample
(µL)
Volume for Ten sample
(µL) with excess of
12.5%
Taqman® PreAmp Master
mix (2X)
12.5 140.62
MegaplexTM PreAmp
Primers (10X)
2.5 28.13
Nuclease-free water 7.5 84.37
Total 22.5 253.12
Invert the master mix six times to mix thoroughly and briefly centrifuge. In a 96-well
MicroAmp® Optical Reaction Plate place, 22.5 µL of the PreAmp master mix into each
appropriate well. Pipette 2.5 µL of the RT reaction product into each appropriate well
brings the final reaction volume to 25 µL. The plate was sealed using MicroAmp® Clear
Adhesive Film and incubated on ice for 5 min. Pre-Amplification parameters were set up
in accordance with the following:
Stage Temp (˚C) Time (min)
Hold 95 10
Hold 55 2
Hold 72 2
Cycle
(12 Cycles)
95 15 sec
60 4
Hold 99.9 10
Hold 4 ∞
2.4.7.3 PCR
RT products were diluted using 75 µL nuclease-free water, inverted to mix and spun down
briefly by centrifugation. The miRNA PCR reaction was carried out using the Taqman®
93
2X Universal PCR Master Mix (No AmpErase® UNG) and in accordance with the below
protocol in a 1.5 mL Eppendorf tube:
Component Volume for One array (µL) with
12.5% excess
Taqman® 2X Universal PCR Master Mix
(No AmpErase® UNG)
450
Diluted PreAmp Product 9
Nuclease-Free Water 441
Total 900
The tube was inverted 6 times to mix and briefly centrifuged. A volume of 100 µL of the
PCR reaction mix was loaded into each filling port of the Taqman MicroRNA Array
(Figure 2.5).
Taqman
microRNA
Array
Filling
Port 1
Filling
Port 2
Filling
Port 3
Filling
Port 4
Filling
Port 5
Filling
Port 6
Filling
Port 7
Filling
Port 8
Multiplex
RT
Primer
Pool
1
2
3
4
5
6
7
8
Plate was centrifuged so that all micro fluidic wells are filled with RT reaction mixture
and then subsequently sealed. Arrays were incubated for 10 min on ice and were loaded
on HT7900 real time PCR machine and ran under the following protocol:
Step AmpliTaq Gold®
Enzyme
Activation
PCR
HOLD CYCLE (40 cycles)
Denatured Anneal/Extend
Time (min) 10 15 sec 60 sec
Temp (˚C) (˚C) 95 95
2.4.7.4 Data Analysis
Data was analysed using Real-Time StatMinerTM software form Integromics.
94
2.4.8 Singleplex qRT-PCR validation: Taqman® microRNA assays
Taqman® microRNA assay kits are pre-formulated primer and probe sets designed to
detect and quantify mature miRNAs in real time. This assay can discriminate mature
miRNAs from their precursor sequences and quantify small amounts (1-10 ng). Assays
range for the majority of content found in the miRBase miRNA sequence repository.
Total RNA was reverse transcribed by specific primers within the miRNA RT kit while
the PCR reaction is performed using another set of miRNA specific primers together with
the Taqman® Universal Master mix (Figure 2.6 A). miRNA PCR Amplicon detection is
based on a DNA probe containing a fluorescent tag (FAMTM) at the 5’ end of the probe
along with a non-fluorescent quencher (NFQ) at the 3’ end of the probe. The Taqman
minor groove binding (MGB) probe hybridizes to complementary sequences within the
forward and reverse primers. This intact probe prior to DNA binding holds the fluorescent
tag in close proximity to the non-fluorescent quencher resulting in no fluorescence (Fig
2.6 B).
Figure 2.6: RT/PCR Mechanism and FAMTM – based detection mechanism
Polymerisation
Strand Displacement
Cleavage
Complete Polymerisation
B) Probe MechanismA) Two-Step RT/PCR Mechanism
95
During DNA polymerization, the DNA polymerase cleave only probes that are bound to
the complementary sequences thus releasing the fluorescent FAMTM tag from the
suppressive proximity of the quencher. As the number of PCR cycles increases the level
of probe hybridization/cleavage/fluorescence increases.
2.4.8.1 Reverse transcription
RT master mix for each sample was prepared separately using the Taqman microRNA
Reverse Transcription kit (ABI, 4366596) from Applied Biosystems.
Component Master Mix (15 µL-Single
Reaction)
100mM dNTPs (with dTTP) 0.15
MultiScribeTM Reverse Transcriptase, 50 U/µL 1.00
10X Reverse Transcriptase Buffer 1.5
5X RT Primer 3
RNase Inhibitor, 20 U/µL 0.19
Nuclease-free water 4.16
Total Volume 10
A separate master mix reaction was prepared without MultiScribeTM RT enzyme for each
primer set to be used within the assay in order to validate genomic contamination of the
primers and reaction components. RT-master mix was mixed gently and centrifuged prior
to distribution. An aliquot of 10 µL of RT-master mix was transferred into a 0.5 mL
Eppendorf and placed on ice until RNA was prepared. RNA samples were pre-diluted to
a working concentration of 2 ng/µL in nuclease-free water. For each 15 µL reaction, 5 µL
of RNA (1-10 ng/reaction) was added to the 10 µL master mix, mixed gentle and
centrifuged. RT was performed using the following parameters in a G-StormTM
thermocycler.
Step Time (min) Temperature (˚C)
HOLD 30 16
HOLD 30 42
HOLD 5 85
HOLD ∞ 4
96
cDNA samples were stored at -20˚C until further required.
2.4.8.2 PCR
PCR reaction for individual miRNA was performed using the Taqman 2X Universal PCR
Master Mix (ABI, 4324018) from Applied Biosystems and was prepared in accordance
with the following table.
Component 20 µL (Single reaction)
Taqman MicroRNA Assay (20X) 1.00
Product from RT reaction (Minimum 1:15 Diltuion 1.33
Taqman 2X Universal PCR Master Mix, No AmpErase®
UNG
10.00
Nuclease-free water 7.67
Total Volume 20
To each well of a MicroAmpTM optical reaction plate, 16 µL of PCR-master mix,
containing appropriate primer, was loaded. From the pre-diluted RT reaction, 5 µL of
product was added to each PCR-master mix. The plate was sealed with microAmpTM
optical adhesive film (ABI, 4311971) and then incubated for 10 min on ice. PCR was
performed using 9600 emulsion mode of thermocycler with reaction volume of 20 µL in
a HT7600 real time PCR machine. Conditions for the PCR reaction were as follows:
Step AmpliTaq Gold®
Enzyme
Activation
PCR
HOLD CYCLE (40 cycles)
Denatured Anneal/Extend
Time (min) 10 15 sec 60 sec
Temp (˚C) (˚C) 95 95
The table below represents all miRNA assay kits used over the course of this Ph.D.
97
Table 2.3: microRNA assay kits (Applied Biosystems)
Applied BiosystemsTM microRNA assay kits
miR Cat. No Assay I.D.
U6 snRNA (Endogenous
control)
4427975 001973
miR-18a 4427975 002422
miR-23b 4427975 000400
miR-23b* 4427975 002126
miR-34a 4427975 000426
2.4.9 MessageAmpTM II aRNA amplification kit
The MessageAmpTM II aRNA amplification kit is an RNA amplification protocol. The
procedure consists of reverse transcription with an oligo(dT) primer bearing a T7
promoter using ArrayScriptTM reverse transcriptase (RT), designed to produce high yields
of first-strand cDNA. ArrayScript RT mediates the synthesis of virtually full-length
cDNA, which is the best way to ensure production of reproducible microarray samples.
The cDNA then undergoes second-strand synthesis and clean-up to become the template
for in vitro transcription (IVT) with T7 RNA polymerase. To maximise aRNA yield,
Ambion MEGAscript® IVT technology is used to generate hundreds to thousands of anti-
sense RNA copies of each mRNA in a sample (aRNA is often referred to as cRNA). RNA
samples of limited amounts (0.1-100 ng) can be put through two rounds of amplification
if required. Figure 2.7 demonstrates the amplification procedure using MessageAmpTM
II aRNA amplification kit.
98
Figure 2.7: MessageAmpTM II aRNA amplification procedure
Figure 2.7: A schematic flow diagram of the procedure undertaken to amplify low quantities of RNA
in preparation of carrying out microarray studies requiring a high amount RNA per card. This
diagram was sourced from http://tools.lifetechnologies.com/content/sfs/manuals/cms_056141.pdf.
Amplification procedure was carried out in accordance with the manufacturer’s
specification (Life Technologies, #AM1751). First strand cDNA synthesis was carried
out on a maximum of 1000 ng of RNA in a maximum volume of 10 µL topped up with
99
high quality water. This mixture was vortexed briefly and centrifuged to collect. The
reverse transcription master mix was prepared at room temperature in nuclease-free
including a 5% excess to cover pipetting errors according to the following cocktail:
Reverse Transcription master mix (for a single 20 µL reaction)
Amount (µL) Component
1 Nuclease-free water
1 T7 Oligo(dT) primer
2 10X First Strand Buffer
4 dNTP Mix
1 RNase Inhibitor
1 ArrayScript
This mixture was gently vortexed and centrifuged briefly to collect and placed on ice until
required. 10 µL of Reverse Transcription Master Mix was added to each RNA sample
forming the final 20 µL reaction volume. The mixture was pipetted up and down 2-3
times, flicked 3-4 times and briefly centrifuged to collect. Samples were placed in a
thermo cycler and subjected to the following conditions:
Temp Time Cycles
42°C (50°C lid) 2 h 1
4°C Store
Samples were placed on ice immediately after first-strand synthesis.
Second-strand cDNA synthesis was carried out next. On ice, the second-strand Master
Mix was prepared in nuclease-free water according to the cocktail below to a final volume
of 80 µl:
Second Strand Master Mix (For a single 100 µL reaction)
Amount (µL) Component
63 Nuclease-free Water
10 10X Second Strand Buffer
4 dNTP Mix
2 DNA polymerase
1 RNase H
This reaction included a 5% excess to cover pipetting errors. Master Mix was mixed
gently by vortexing and centrifuged to collect and placed on ice. 80 µL of second strand
master mix was transferred to each 20 µL first-strand sample. Mix was pipetted up and
100
down 2-3 times, then flicked 3-4 times and centrifuged to collect. Samples were placed
on a thermo cycler in accordance with the following steps:
Temp Time Cycles
16°C (no heated lid) 2 h 1
4°C hold
Be sure to deactivate the heated lid and allow thermo cycler to cool to 16°C before
incubation as temperatures above this can compromise aRNA yield. Samples can be
placed on ice for no more than 1 h or frozen at -20°C.
Double-stranded cDNA was then purified. 250 µL of cDNA binding buffer was added to
each sample, mixed thoroughly by pipetting up and down 2-3 times, flicked 3-4 times
followed by a quick spin to collect. The entire sample was placed onto the centre of the
cDNA filter cartridge. Samples were centrifuged for 1 min at 10,000 g or until mixture
was fully through the filter. The flow-through was discarded and the cDNA filter cartridge
was replaced into the collection tube. 500 µL of wash buffer was added to each cDNA
filter cartridge. Samples were centrifuged for 1 min at 10,000 g. The flow-through was
discarded and the cDNA filter cartridge was centrifuged for an additional minute to
remove trace amounts. cDNA filter cartridge was transferred to a clean elution tube. 18
µL of nuclease-free water pre-incubated at 50-55°C was added to the centre of each
cDNA filter cartridge. This was left to incubate for 2 min and then centrifuged for 1 min
at 10,000 g. The remaining ~16 µL contains double-stranded cDNA.
In vitro transcription to synthesise aRNA was carried out using un-modified or biotin-
labelled UTP. The 16 µL of double-stranded cDNA prepared previously was transferred
to a PCR tube. IVT master mix was prepared for a maximum reaction volume for 40 µL
to achieve highest aRNA yield in accordance with the following cocktail:
101
Un-modified 40 µL reaction Component
16 µL Double-stranded cDNA: previous step
IVT Master Mix for a single reaction
4 µL T7 ATP
4 µL T7 CTP
4 µL T7 GTP
4 µL T7 UTP
-- Biotin-11-UTP, 75 mM
4 µL T7 10X Reaction Buffer
4 µL T7 Enzyme Miz
Master Mix was mixed well by vortexing and centrifuged to collect the IVT master mix
at the bottom of the tube and placed on ice. Transfer IVT master mix to each sample and
thoroughly mix by pipetting 2-3 times, flicking 3-4 times with a brief centrifugation to
collect. Once mixed, samples were placed in a thermo cycler and processed at the
following conditions:
Temp Time Cycles
37°C (100-105°C heated
lid)
14 h 1
4°C hold
Reaction was stopped by adding Nuclease-free water to each aRNA sample to bring the
final volume to 100 µL, mixed by vortexing and stored at -20°C.
aRNA purification was final carried out to remove enzymes, salts, and unincorporated
nucleotides from the aRNA. 200 µL of nuclease-free water for each samples was pre-
heated to 55°C. For each sample, an aRNA filter cartridge was placed into an aRNA
collection tube. 350 µL of aRNA binding buffer was added to the 100 µL of aRNA. 250
µL of 100% ethanol was added to each aRNA sample and mixed by pipetting up and
down 3 times-DO NOT VORTEX OR CENTRIFUGE. Pipet each sample onto the
centre of the filter in the aRNA filter cartridge/collection tube. Centrifuge for 1 min at
10,000 g. Discard flow-through and replace the aRNA filter cartridge back into aRNA
collection tube. Apply 650 µL wash buffer to each aRNA filter cartridge. Centrifuge for
1 min at 10,000 g. Discard flow-through and spin aRNA filter cartridge for an additional
minute to remove trace amounts of wash buffer. Transfer filter cartridge to a fresh aRNA
collection tube. To the centre of the filter, add 200 µLL pre-heated nuclease-free water
and incubate samples on a heat block at 55°C for 10 min. Centrifuge for 1.5 min at 10,000
102
g and quantify. 33.3 ng/µL of aRNA is required for microarray processing or a second
round of amplification can be performed. Samples are stored at -80°C for up to one year.
2.4.10 CHO-specific oligonucleotide arrays
Following from the amplification using the MessageAmpTM aRNA amplification kit (see
2.4.9) of the low quantity RNA generated from the biotin-labelled miRNA pull-down,
100 ng of total RNA from each sample underwent cDNA synthesis, followed by clean-
up, overnight IVT amplification and labelling using the GeneChip 3’ IVT Express Kit
(#901229, Affymetrix) according to manufacturer’s specifications. cRNA clean-up was
carried out using the GeneChip Sample Clean-up Module (#900371, Affymetrix)
according to manufacturer’s specifications. The custom CHO oligonucleotide
WyeHamster3a microarray (Affymetrix) used in this target identification analysis
contains a total of 19,809 CHO-specific transcripts, combining library-derived and
publicly-available hamster sequences.
2.4.11 Gel and LB media/LB agar preparation
2.4.11.1 Agarose gel preparation
Depending on the percentage gel desired, agarose (Sigma-Aldrich, A9539) was added to
1X TAE (Thermo, #B49) with 2% (w/v) being considered a high percentage for resolution
of small DNA bands (~ 100 bp). The solution wass heated to boiling until completely
dissolved and cooled under a running cold tap. 0.5 µg/mL of ethidium bromide was added
to the gel before pouring. Prior to pouring the gel, gel try was sealed at both ends using
autoclave tape and left to set in the dark upon pouring.
2.4.11.2 LB media preparation
For 1 litre of LB media, 10 grams of Tryptone (Sigma-Aldrich, #T7293), 5 grams of Yeast
(Sigma-Aldrich, #92144) and 10 grams of NaCl (Sigma-Aldrich, S9888) was dissolved
in 1000 mL of UHP and subsequently autoclaved.
103
2.4.11.3 LB agar plate preparation
LB agar preparation follows the same ingredients as LB media with the addition of 15
grams of agar select (Sigma-Aldrich, A5306) in 1,000 mL of UHP and subsequently
autoclaved. LB agar is then melted in the microwave and allowed to cool at room
temperature until bottle is cool enough to handle. It is at this point that antibiotics are to
be added to the LB agar media mixed well and poured into petri dishes.
2.4.12 Restriction Enzyme Digestion
Restriction enzymes are enzymes isolated from bacteria that are used as tools in molecular
cloning techniques to recognise and cut DNA at specific genetic locations to produce
DNA fragments known as restriction fragments. Restriction enzymes are used to create
acceptor sites or “sticky ends” within multiple cloning sites (MCS) of plasmid backbones
and the recombinant DNA insert.
Digestion reactions were set up in 0.5 mL PCR tubes at a reaction volume of 20 µL. The
reaction volume can vary depending on the amount of DNA required. The following
represents the components required for the digestion of 2µg of DNA with a single
restriction enzyme in a volume of 20µL.
Component Volume (µL)
Plasmid DNA 2 µg
NEBuffer 1-4 (10X) 2
BSA (100X) 0.2
Restriction enzyme 1U (0.5-1 µL)
Nuclease-Free water To 20 µL reaction volume
Enzymatic digestions are carried out at a temperature of 37 ˚C on a hot plate until
complete digestion is achieved. 1 Unit of a restriction enzyme is defined as the digestion
of 1 µg of λ DNA in 1 h at a total volume of 50 µL (see NEB website
https://www.neb.com/). Double digests can be carried out within the same reaction vessel
depending on the compatibility and activity of the respective enzymes in the reaction
buffers. Enzymatic activity charts can be found on the New England Biolabs website cited
above.
104
2.4.13 Enzyme modification of linearized plasmid DNA
2.4.13.1 Alkaline Phosphatase (AP)
Alkaline Phosphatase treatment is exploited in molecular cloning as a technique to
remove the 5’-phosphate residue from the DNA backbone after the circularised plasmid
has been linearized, producing either 3’ or 5’ over hangs with exposed hydroxyl or
phosphate groups, respectively. The phosphate group on C-6 and the hydroxyl (OH)
group on C-3 serve as the basis of the chemical reaction of polymerisation of the DNA
strand through a process known as condensation.
The purpose and function of CIP treatment in molecular cloning is to remove the 5’-
phospate group of the DNA backbone inhibiting self-re-ligation of the linearized plasmid
unless in the presence of DNA ligase. CIP treatment can be used on the linearized plasmid
vector whether the exposed termini are incompatible, overhangs are not complementary,
or compatible, blunt ends, but is more commonly used in the latter.
Alkaline Phosphatase Protocol
Prior to AP treatment, during the elution step of cleaning up the restriction digest product,
elution was carried out using 45µL of elution buffer (EB). 4.5 µL of AP Buffer (NEBuffer
3 (50 mM Tris-HCl, 100 mM NaCl, 10 mM MgCl2, 1 mM Dithiothreitol, pH 7.9 @ 25˚C).
0.4 µL of Alkaline Phosphatase enzyme (Storage Conditions: 10 mM Tris-HCl, 50 mM
KCl, 1 mM MgCl2, 0.1 mM ZnCl2, 50% Glycerol, pH 8.2 @ 25˚C) was added to the
reaction vessel and mixed gently by using a pipette. Reaction was incubated on a hotplate
at 37 ˚C for 1 h. 45 µL of Chloroform (SIGMA) was added to the reaction mixture
subsequent to the incubation period and mixed rigorously by vortexing for 1 minute.
Centrifugation at 13,000 rpm for 4 min was performed to separate both layers. The top
layer was transferred to a fresh clean Eppendorf tube without any contamination from the
lower layer. To ensure maximum amount of CIP treated DNA, 10 µL of nuclease-free
water was added to the previous tube containing the chloroform and centrifuged again at
13,000 rpm for a further 2 min. Using a lower volume pipette, the top layer was removed
again and placed into the tube containing the top layer from the previous centrifugation
105
step. DNA was cleaned using the PCR clean-up kit (see 2.6.5) and stored at -20˚C until
further required.
2.4.13.2 Kinase Treatment
T4 PNK catalyses the transfer of the terminal gamma (γ) phosphate group from
Adenosine Tri-phosphate (ATP) to the 5’-hydroxyl termini of DNA and RNA (Fig 2.10).
This enzyme can be used for the phosphorylation of the 3’-ends of RNA to prevent
circularisation, phosphorylate the 5’ ends of DNA and RNA in order to provide the
necessary phosphate backbone required for ligation processes and label 5’-hydorxyl ends
of DNA with [γ-32P]-ATP (Hui Zhu et al, 2007). In some instances, with the presence of
an excess of Adenosine Di-Phosphate, the reverse reaction of T4 PNK transfers the 5’-
phosphate from DNA/RNA to ADP thereby fulfilling the role of (CIP) Calf Intestinal
Phosphatase which prevents linearized plasmids from circularising.
0.5 µL of both sense and anti-sense strands that were previously annealed were added to
a 0.5 mL PCR tube. 2 µL of 10 x T4 kinase buffer (500 mM Tris-HCl, 100 mM MgCl2,
1 mM EDTA, 50 mM dithiothreithol, 1 mM spermidine pH 8.2 (at 25˚C). 0.4 µL of 50
mM ATP (pH 7.5) was added to this reaction as the source of phosphate. This reaction
was made up to a volume of 20 µL with the addition of 17 µL nuclease free water
(Ambion® Cat. AM9932). Lastly 0.4 µL (4U) of Kinase enzyme (ROCHE, storage
buffer, 50 mM Tris-HCl, 1 mM dithiothreithol, 0.1 mM EDTA, 1 µM ATP, 50% glycerol
(v/v), pH 7.5 (at 25˚C) was added. Using a heating block this reaction was incubated at
37 ˚C for 30 min. Upon completion of the 30 minute incubation, the kinase treated
linearized DNA was ready for ligation and can be used up to one year later and must be
stored at -20˚C. The T4 PNK supplied by ROCHE must not undergo the 20 minute
incubation at 65˚C like the SIGMA supplied enzyme.
2.4.13.3 Klenow Treatment
Escherichia Coli DNA Polymerase I: All DNA polymerase enzymes add
deoxyribonucleotides to the 3’-hydroxyl terminus of a primed double stranded DNA
molecule. Synthesis is strictly in a 5’-3’ direction. Many DNA polymerases have a 3’-5’
106
exonuclease activity associated with the polymerase activity. In the absence of dNTPs,
this activity will catalyze the degradation from a free 3’hydroxyl end of both single- and
double-stranded DNA. In the presence of dNTPs, the exonuclease activity on double-
stranded DNA is inhibited by the polymerase activity. Some DNA polymerase enzymes
(eg., E. Coli DNA polymerase I) have an associated 5’-3’ exonuclease activity. This
activity degrades double-stranded DNA from a free 5’-hydroxyl end. This allows Nick
Translation which degrades in the 5’-3’ direction ahead of the polymerizing activity.
Klenow treatment can be used to trim overhangs to make them compatible with other
restriction enzymes where combinations are not applicable. Additionally, the klenow
treatment can also inhibit self-re-ligation (Fig. 2.8).
Klenow Protocol
Upon digestion and cleaning of the plasmid/DNA fragment of interest, 1 µg of digested
product was loaded into a 0.5 mL PCR tube. For a 20 µL reaction volume, 2 µL of 10X
NEBuffer 2 (10 mM Tris-HCl, 50 mM NaCl, 10 mM MgCl2, 1 mM Dithiothreitol (pH
7.9 at 25˚C)) is added to the reaction vessel. See table below for variations according to
final reaction volume. dNTPs were added to a final concentration of 33 µM. 1U of
Klenow enzyme (Storage Conditions: 25 mM Tris-HCl, 1 mM Dithiothreitol, 0.1 mM
EDTA, 50% Glycerol (Ph 7.9 at 25˚C)) is added for every 1 µg of DNA in the reaction.
Reaction volume is made up using Nuclease-free water. Reaction mixture should be
mixed gently using a pipette before placing on a heating block at 25˚C for 15 min. Klenow
reaction was stopped by the addition of 10mM EDTA and heating at 75˚C for 20 min.
Reaction
Vol. (µL)
Template
DNA (µg)
1xNEBuffer
2 (µL)
Klenow dNTPs (33
µM) (µL)
H2O
20 1 µg 2 0.1-0.5 (1-
5U)
0.5 Up to 20
µL
30 1 µg 3 0.1-0.5 (1-
5U)
0.5 Up to 30
µL
50 1 µg 5 0.1-0.5 (1-
5U)
0.5 Up to 50
µL
107
Figure 2.8: Cloning process and Klenow treatment
2.4.14 Gel extraction
Plasmid DNA fragments of interest containing gene coding sequences were extracted
from DNA agarose gels using the QIAGEN® QIAquick Gel Extraction kit (QIAGEN®,
28704). The DNA fragment was excised from a low-melting agarose gel using a clean
108
scalpel. 400 mg of agarose can be processed by the spin column so it is important to
minimize the size of the gel slice by removing extra agarose. The gel slice was finely
chopped up and placed into a microcentrifuge tube. The gel slice was then weighed by
weighing an empty microcentrifuge tube and the one containing the gel slice. 3 volumes
of Buffer QG was added to 1 volume of gel (100 mg ~ 100 µL). This mixture was
incubated at 50˚C for 10 min and vortexed every 2-3 min until the gel slice was
completely dissolved. Once the gel was dissolved, the mixture colour was yellow. If the
mixture was orange or violet then 10 µL of 3M sodium acetate (Sigma-Aldrich,
#W302406) was added allowing efficient adsorption of DNA to the QIAquick membrane.
1 gel volume of isopropanol (Sigma-Aldrich, #190764) was added to the sample and
mixed using a pipette. A QIAquick spin column was placed into a 2 mL collection tube.
To bind the DNA, the sample mixture was applied to the QIAquick spin column and
centrifuged for 1 minute at 13,000 rpm. The flow through was discarded and the
QIAquick column was placed back into the collection tube. The remaining traces of
agarose were washed from the column by loading 500 µL of Buffer QG to the QIAquick
column and spinning at 13,000 rpm for 1 minute. A second wash step was performed by
applying 750 µL of Buffer PE to the QIAquick column followed by centrifugation at
13,000 rpm for 1 minute. The flow through was discarded and a second spin for 1 minute
was performed to dry the column. DNA elution was carried out by placing the QIAquick
column into a fresh microcentrifuge tube and adding 30-50 µL of Buffer EB or water
directly onto the QIAquick membrane and left standing for 1 minute at room temperature
before centrifugation at 13,000 rpm for 1 minute. Eluted DNA was quantified by
measuring the OD260nm using a NanoDrop spectrophotometer. Plasmid was sorted at -
20˚C for further use.
2.4.15 PCR Purification
All plasmid DNA was purified using the QIAGEN® QIAquick PCR Purification kit
(QIAGEN®, 28104) in accordance with the manufacturer’s specifications. This clean up
kit was necessary for the purification of single-/double-stranded DNA from PCR and all
other enzymatic reactions such as Kinase/CIP/Klenow. Into the PCR mix, 5 volumes of
Buffer PB was added and mixed by pipetting. The colour of this mix will turn yellow if a
pH indicator has been added to Buffer PB. If this solution appears orange or violet then
109
10 µL of 3M Sodium Acetate is added and mixed. A QIAquick spin column was placed
into a 2 mL collection tube. In order to bind the DNA, the sample mix was applied to the
QIAquick column and spun by centrifugation at 13,000 rpm for 1 minute. The flow
though was discarded and the QIAquick column was placed back into the same 2 mL
collection tube. A wash step was performed by adding 750 µL of Buffer PE to the
QIAquick column and centrifuged at 13,000 rpm for 1 minute. Flow through was
discarded and the QIAquick column was placed back into the collection tube and spun
for a further 1 minute at 13,000 rpm to dry the membrane. Before elution of the purified
DNA, the QIAquick spin column was placed into a clean microcentrifuge tube. DNA was
eluted by adding 30-50 µL of Buffer EB or water directly to the QIAquick membrane and
allowing sitting for 1 minute at room temperature followed by spinning at 13,000 rpm for
1 minute. Quantification of DNA was determined using a NanoDrop spectrophotometer
and measuring the OD260nm. Purified DNA was stored at -20˚C until required.
2.4.16 Ligation
The T4 DNA ligase kit from ROCHE® (Cat. No. 10 481 220 001) was used for all ligation
reactions carried out. It is recommend that no more than a total of 1 µg of DNA is to be
used in any given ligation reaction and a range of 1-5 Units of T4 DNA ligase enzyme.
The online tool LIGATION CALCULATOR (http://www.insilico.uni-
duesseldorf.de/Lig_Input.html) was used to determine the optimum ratio in nanograms
of insert to vector backbone (Fig 2.9). To insure efficient ligation it is important to have
a saturation of insert to vector by using the recommended ratios of 1:3 or 1:5.
110
Figure 2.9: Online LIGATION CALCULATOR tool
Figure 2.9: Online LIGATION CALCULATOR tool: All relevant information is to be entered into
the above fields including vector backbone size (after restriction enzyme digestion) in base pair (bp),
vector amount in nanograms (ng) after clean-up, insert size in bp and finally the ration of insert to
vector (1:3 or 1:5).
A range of ~80-150 ng of vector backbone was used in all ligation protocols depending
on the yields achieved after the previous two cleaning and CIP treatments beforehand.
The concentration of insert used was determined based on the fragment size and the
amount/size of vector being used (Ligation Calculator). All reactions were carried out at
20 µL but scale up as far as 50 µL can be done if required. 2 µL of 10x T4 DNA ligase
reaction buffer (660 mM Tris-HCl, 50 mM MgCl2, 10 mM ATP, 50 mM EDTA, pH 7.5
at 25˚C) was added to the reaction mixture. Nuclease-free water was used to make up the
reaction volume to 19 µL prior to the addition of the T4 DNA ligase enzyme. Finally 1
µL (1 U) of T4 DNA ligase enzyme (20 mM Tris-HCl, 60 mM KCl, 5 mM
Dithioerythritol, 1 mM EDTA, 50% Glycerol, pH 7.5 at 4˚C) was added to the reaction
mixture creating a reaction volume of 20 µL. This reaction was incubated overnight on
ice water. This allows the ligation reaction to commence over a temperature gradient.
Ligated samples were stored at -20˚C. It is important to set up an identical reaction
containing all components except the fragment to be inserted (no insert control). This
accounts for the frequency of self-ligation of the DNA backbone and incomplete digestion
of the vector backbone resulting in non-positive colonies.
111
2.4.17 Transformation of competent cells
Two kits supplied by InitrogenTM were used for the transformation of ligated vectors or
for the amplification of plasmid vector stocks. These kits were the Subcloning
EfficiencyTM DH5αTM Competent cells (Sigma-Aldrich, Cat. No.18265-017) and MAX
Efficiency® DH5αTM Competent Cells (Sigma-Aldrich, Cat No. 18258-012). The only
difference between the two kits above is that the MAX Efficiency® DH5αTM Competent
Cells kit has a greater transformation efficiency of 1 x 103 transformants/µg pUC19 DNA.
A 14 mL polypropylene Round-Bottom tube (Falcon®) was placed on ice while the water
bath (NICKLE-ELECTRO Ltd., Clifton, 14 Litre Unstirred Thermostatic Bath, series
NEI-14) was pre-heated to 42˚C. 50 µL of competent cells were transferred into each tube
and kept on ice. ~5 µL of DNA (Ligation product, less for stock plasmid) was added to
the competent cells and tapped to mix. This mixture was left on ice for ~20 min. The
competent cells/DNA was heat-shocked by placing in the water bath at 42˚C for 45
seconds. This step serves to cause contraction of the membrane pores resulting in the
uptake of DNA from the surrounding environment. Upon heat shock the cells are then
immediately placed on ice for a further 3 min which serves to constrict the membrane
pore once again. 500 µL of pre-heated (37 ˚C) S.O.C media (supplied in MAX
Efficiency® DH5αTM Competent Cells kit) was added to the transformed competent cells.
Cells were incubated for ~1 h at 37 ˚C on a shaker platform. After an h, the contents were
transferred into a 1.7 mL microcentrifuge tube (Costar®, Corning Incorporated, Cat. No.
3620) and centrifuged at 3000 rpm for 4 min. All but 50 µL of SOC media was decanted
from the bacterial pellet. The pellet was resuspended using a pipette and transferred onto
a LB agar plate containing the appropriate selection antibiotic (Section 2.4.11). Using a
sterile microbiological spreader (Copan Innovation, Cat. No. 22-091-241), the
transformed competent cells were homogenously spread over the LB agar plate
containing the appropriate antibiotic selecting agent. These plates were then sealed with
Parafilm® M (SPI Supplies®, Cat. No. 01851-AB) and incubated at 37 ˚C for 16-24 hs.
112
Figure 2.10: Test and Control plate under Ampicillin selection
Figure 2.10: The requirement of carrying out a no-insert control during the ligation process is
demonstrated. A) Colonies as a result of successful insertion causing circularisation in addition to
circularised plasmids with no insert as a result of incomplete digestion. B) Colonies formed only as a
result of incomplete digestion giving an indication of colony number without target inserts.
After a time period of 16-24 hs of incubation, competent cells that successfully took up
the plasmid will form single round colonies on the agar plates. The presence of the
antibiotic selecting agent will select for colonies efficiently expressing the functional
circularised plasmid that harbours the antibiotic resistance gene. Figure 2.10 A and B
shows a test and control plate under ampicillin (Sigma-Aldrich, #A9393) selection after
24 hs of growth. The no-insert control plate was important to account for occurrence of
incomplete digestion and random circularisation of the vector backbone not containing
the fragment of interest. By counting the number of colonies on the test and control plate,
it gives an idea of the number of colonies that is required to be picked and mini-prepped
(Section 2.6.8) that will contain the plasmid of interest.
2.4.18 DNA mini-prep of plasmid DNA
The QIAGEN QIAprep Spin Miniprep kit (QIAGEN, 27104) was used according to the
manufacturer’s instructions for the isolation of plasmid DNA from a 5mL culture. This.
After incubation, cells were pelleted by centrifugation at 4000 rpm for 10 min. Cell pellets
A: Test Plate B: No-Insert control Plate
113
were resuspended in 250 µL of buffer P1 and transferred to a microcentrifuge tube. 250
µL of buffer P2 was added to this solution and mixed gently by inversion. A 350 µL
volume of buffer N3 was subsequently added to this solution and again mixed gently by
inversion. This mixture was centrifuged for 10 min at 13,000 rpm. The decanted
supernatant was applied to the QIAprep spin column and centrifuged for 60 seconds at
13,000 rpm. This flow though was discarded. The column was washed by adding 500 µL
of Buffer PB and centrifugation for a further 60 seconds at 13,000 rpm. Again, the flow
though was discarded. A second wash was performed using 750 µL of Buffer PE and
centrifuging for 60 seconds. The flow through was discarded and the column was dried
by an additional centrifugation step for 60 seconds to remove any remaining wash buffer.
The spin column was then placed into a clean microcentrifuge tube. The DNA was eluted
by adding 30-50 µL (depending on the desired DNA yield) of Buffer EB or water to the
spin column. This was allowed to stand for 1 minute at room temperature (RT) before
centrifugation at 13,000 rpm for 1 minute. The eluted DNA was reapplied to the spin
column and centrifuged for an additional 60 seconds to collect any remaining plasmid
DNA. Quantification of the eluted plasmid DNA was determined using a NanoDrop
spectrophotometer by measuring at OD260nm. Plasmid was stored at -20˚C until required.
2.4.19 DNA midi/Maxi-prep of plasmid DNA
Stock plasmids were amplified and purified using an Endotoxin-free Plasmid DNA
Purification Midi/Maxi kit (Nucleobond® Xtra Midi plus EF, Macherey-Nagel,
740416.50). Typical yield from the Nucleobond® Xtra Midi kit is 250 µg of plasmid DNA
compared to the 1000 µg from the Nucleobond® Xtra Maxi kit. Both midi and maxi preps
were carried out according to the manufacturer’s specifications with midi prep being the
general route of plasmid amplification. 5 mL of inoculated LB media was added to 100
mL of pre-warmed LB media once again under the selection of antibiotic and incubated
over night at 37 ˚C on a shaker platform. Pellets were harvested by transferring this 100
mL to a 250 mL centrifuge flask and spun at 6,000 g for 10 min at 4˚C. Supernatants were
discarded completely and pellets were either processed immediately or stored at -20˚C
until required. The Bacterial pellet was resuspended in 8 mL of Buffer RES-EF by
pipetting up and down until no clumps remained and transferred into a 50 mL Falcon
tube. Cells were lysed open by adding 8 mL of Buffer LYS-EF and mixed gently by
114
inverting. This mixture was incubated at room temperature (18-25˚C) for 5 min. During
this incubation period, the Nucleobond® Xtra Column with the inserted column filter was
equilibrated with the addition of 15 mL Buffer EQU-EF around the rim of the inserted
column filter. The cell lysate mixture was neutralised with the addition of 8 mL Buffer
NEU-EF, inverted 10-15 times and incubated on ice for 5 min. The crude lysate was
homogenised by inverting before applying it directly to the equilibrated Nucleobond®
Xtra Column Filter and allowed to empty by gravity flow. The first wash consisted of the
addition of 5 mL Buffer FIL-EF to the rim of the NucleoBond® Xtra Column Filter and
once again left to empty by gravity flow. Once the column was emptied completely, the
NucleoBond® Xtra Column Filter was removed and discarded. The second wash step
washes the NucleoBond® Xtra Column by adding 35 mL of Buffer ENDO-EF directly
into the column and allowing emptying by gravity flow. The third and final wash requires
the addition of 15 mL Buffer WASH-EF to the column and emptying by gravity flow.
Plasmid DNA was eluted into a 20 mL Universal by adding 5 mL of Buffer-ELU-EF to
the column. For elution of DNA using the NucleoBond® Finalizer the DNA was
precipitated with the addition of 3.5 mL isopropanol (SIGMA), vortexed well and left to
sit for 2 min. The plunger was removed from a 30 mL Syringe and the NucleoBond®
Finalizer was attached. The precipitate was loaded into the syringe and the plunger was
reinserted into the syringe. Using constant force the precipitate passes through the
NucleoBond® Finalizer and the flow though is discarded. The Finalizer was removed
before the removal of the plunger and then reattached. Using the same technique a wash
step with 2 mL Endotoxin-free 70% EtOH was carried out again discarding the flow
though. The filter membrane of the NucleoBond® Finalizer was dried by removing the
plunger and reinserting it into the syringe and expelling the air though the filter. This was
performed three times. Finally, the plasmid DNA was eluted in 400 µL of Buffer TE-EF
or H2O-EF by using a 1 mL syringe attached to the NucleoBond® Finalizer. The
concentration of eluted plasmid DNA was determined using a NanoDrop
spectrophotometer by measuring the OD260nm. Plasmid was stored at -20˚C until required.
2.4.20 Plasmid diagnostics
Plasmid diagnostics requires a combination of the above listed protocols. Restriction
enzymes are selected that will cut the target insert out of the plasmid constructed in
115
addition to select an enzyme that cuts once at a sequence within the insert. Digestions are
then ran out on an agarose gel stained with ethidium bromide, including a DNA ladder
(Life Technologies, 1 kb Plus DNA Ladder, #10787-018), and compare band sizes to an
appropriate plasmid reference map.
2.5 Mitochondrial studies
2.5.1 Mitotracker® Deep Red FM
Mitochondrial content was assessed using the fluorescent dye Mitotracker® Deep Red
MTDR) FM (InvitrogenTM, M22426). Mitochondrial abundance was determined for 1 x
106 cells stained with a working concentration of 50 ng. Cells were cultured for 72 hs,
pelleted and resuspended in a 1 mL volume of culture medium containing the desired
concentration of MTDR dye and incubated at 37 ˚C for 30 min in a shaker incubator.
Cells were pelleted and resuspended in fresh culture media and analyzed using a FACS
flow cytometer (BD). The MTDR emission wavelength of 665 nm did not interfere with
the emission peak of GFP at 509 nm eliminating the need of compensation in the case of
stable CHO cell lines expressing the miR-sponge construct.
2.5.2 OROBOROS Oxygraph-2k
The OROBOROS/Oxgraph-2k, a two chambered respirometer equipped with a peltier
thermostat and integrated electromagnetic stirrers, was used to determine respirometric
measurements. The oxygen concentration was recorded using the software DatLab
(OROBOROS instruments) and specific oxygen consumption rates were represented as
(pmol O2/s/106 cells). Measurements were performed using 1 x 106 cells/mL in 2 mL of
CHO-S-SFMII for intact cells or in the case of the permeabilised assay, respiration buffer
Mir05 (Hirsch et al. 2012) at 37 ˚C with continuous stirring. The following substrate-
uncoupler-inhibitor-titration (SUIT) protocol was applied in both instances.
In the case of intact cells, oxygen consumption was allowed to stabilise upon inoculation
and sealing of the chambers which resulted in the acquisition of routine respiration (R).
Leak respiration (L) was established upon inhibition of ATP synthase through the
116
addition of oligomycin (2 µg/mL). Maximum respiratory capacity in an uncoupled state
was achieved through addition of FCCP (0.2 µM).
For the more in depth permeabilised protocol, upon initial stabilisation of oxygen
consumption, the cell membrane was permeabilised with 3 µL digitonin (30 µg/106 cells).
Complex I dependent respiration was driven with the addition of glutamate (10 mM) and
malate (0.5 mM). OXPHOS capacity driven by Complex I was achieved by the addition
of ADP (1 mM). Cytochrome C (10 µM) was added to test for aberrations in the integrity
of the outer mitochondrial membrane. Maximal coupled respiration of Complexes I and
II was stimulated by the addition of a Complex II-specific substrate, succinate (10 mM).
By uncoupling the mitochondrial membrane and interfering with the proton gradient,
maximum capacity of the electron transport system (ETS) was obtained using FCCP (0.2
µM). Inhibition of Complex I with rotenone (0.5 µM) shows the maximal uncoupled
respiration via Complex II. Complex II and III were inhibited by the addition of Malonic
acid (5 mM) and Antimycin A (2.5 µM) to estimate the residual oxygen consumption.
Finally, Complex IV was stimulated by the addition of the artificial electron donor TMPD
(0.5 mM) and its reducing agent ascorbate (2 mM) with Complex IV inhibition achieved
through the addition of Azide (100 mM).
2.5.3 Mitochondrial Isolation
Intact mitochondria were isolated using a Mitochondrial Isolation kit for cultured cells
(Thermo Scientific, #88794) in accordance with the manufacturers’ specifications. A
maximum of 6 samples were processed at any one time with 2 x 107 cells being the
processing limits. Following the reagent-based method, 2 x 107 cells were pelleted by
centrifugation ~850 g for 2 min and supernatant was discarded. 800 µL of Mitochondria
Isolation Reagent A was added and vortexed at medium speed for 5 seconds followed by
a 2 minute incubation on ice. 10 µL of Mitochondria Isolation Reagent B was added and
vortexed at maximum for 5 seconds followed by incubation on ice for 5 min with
vortexing at maximum every minute. 800 µL of Mitochondria Isolation Reagent C was
added followed by inversion of samples. Samples were centrifuged at 700 x g for 10 min
at 4°C. Supernatant was transferred to a new 2 mL tube and centrifuged at 12,000 x g for
15 min at 4°C. This contains mitochondria, lysosome and peroxisomes (3,000 x g for 15
min at 4°C to further reduce by >50%). Transfer the supernatant (cytosol fraction) to a
117
new tube. The pellet contains the isolated mitochondria. 500 µL Mitochondria Isolation
Reagent C was added to the pellet and centrifuged at 12,000 x g for 5 min. The supernatant
was discarded. Pellet was maintained on ice until required. Freeze-thawing may
compromise mitochondria integrity.
2.6 Proteomic analysis
2.6.1 Bradford assay
A BSA standard was serially diluted using UHP from a 2 mg/mL stock of Quick startTM
Bovine Serum albumin Standard (Bio-Rad, #500-0206) starting with 1 mg/mL and
spanning 0.5, 0.25, 0.125 and 0 mg/mL. Standards and samples were measured in
triplicates at least by pipetting 5 µl of each into the corner of the base of a well in a 96-
well plate. 250 µL of Quick startTM Bradford 1X Dye Reagent (Bio-Rad, #500-0205) to
each well and incubate in the dark for 10-15 min. Absorbance readings are measured on
a bench top spectrophotometer at a wavelength of 595 nm.
2.6.2 Western Blot
Proteins for analysis by Western Blotting were resolved using SDS-polyacrylamide gel
electrophoresis (SDS-PAGE). Lysis buffer containing urea was added to cell pellet. Cell
lysis was carried out over 20 min on ice and following vortex, the lysate was spun at 4°C
for 15 min at maximum speed to remove insoluble debris. Protein was quantified the
Bradford protocol. Protein samples were diluted in 2X Laemlli buffer (which contains
loading dye, Sigma-Aldrich, S#3401) and appropriate volumes of lysis buffer. Before
loading into the gel, samples were boiled for 5 min and cooled on ice. 5-20 µg of protein
and the molecular weight marker (New England Biolabs) were loaded onto a 4-12%
NuPAGE Bis Tris precast gradient gel (Life Technolgies, NP0322BOX). Electrophoretic
transfer, blocking and development of western blots was carried out. Membranes were
probed with the appropriate antibody of choice diluted in Tris-buffer saline containing
0.1%-20% Tween (TBS-T). An anti-mouse GAPDH monoclonal antibody (Abcam,
#ab8245) was used as an internal control.
118
2.6.3 2-D clean up kit
Lysed protein samples were cleaned up using the ReadyPrepTM 2-D Cleanup Kit (Bio-
Rad, #163-2130) as means of removing any contaminating substances like ionic
detergents, salts, nucleic acids and lipids as a result of the lysis buffer or residual cellular
debris. This will reduce streaking, background staining and other artifacts associated with
substances contaminating 2D/IEF samples. A maximum of 500 µg can be cleaned using
this kit for any one sample. Sample clean-up volume can be up to a maximum of 100 µL
in a 1.5 mL microcentrifuge tube with maintenance on ice unless otherwise indicated.
Always place the centrifuge in the same orientation in the centrifuge.
1-500 µg of protein was transferred to a final volume of 100 µL of MS-grade water. 300
µL of precipitating agent 1 was added to the protein sample and mixed by vortexing with
a 15 minute incubation ice. A further 300 µL of precipitating agent 2 was added to the
mixture and mixed well by vortexing. This was centrifuged at 12,000 g for 5 min to pellet
removing tube promptly after spin step. Without disturbing the pellet, supernatant was
removed using a pipette and discarded. Tube was centrifuged for an additional 15-30
seconds to collect any residual liquid and carefully removed and discarded. 40 µL of wash
reagent 1 was added on top of the pellet and centrifuged at 12,000 g for 5 min, this was
subsequently removed and discarded using a pipette. 25 µL of ReadyPrep proteomic
grade water of UHP was added on top of the pellet and vortexed for 10-20 seconds. 1 mL
of wash reagent 2 (pre-chilled at -20°C for at least 1 h) and 5 µL of wash 2 additive was
added and vortexed for 1 min to mix. Samples were incubated at -20°C overnight. After
this incubation, tubes were centrifuged at 12,000 g for 5 min to form a tight pellet.
Supernatant was removed and discarded with an additional centrifuge for 15-30 seconds
to collect remaining wash. Pellet was air dried at room temperature for no more than 5
min until the pellet appears translucent. Pellet was resuspended by adding an appropriate
volume of 2-D rehydration/sample buffer and vortexed for at least 30 seconds. Tube was
incubated at room temperature for 3-5 min, vortexed again and pipetted up and down a
few times to fully resuspend. Sample was centrifuged to collect at 12,000 g for 2-5 min.
Protein is ready for subsequent analysis or treatment and stored long term at -80°C.
119
2.6.4 Label-free sample digestion (Trypsin and Lys-c)
ProteaseMAXTM Sufactant, Tryspin Enhancer (Promega, #V2072) was used to carry out
and enhance in-solution digestions in accordance with the manufacturer’s specifications.
ProteaseMAXTM Surfactant solubilizes proteins including difficult proteins (membrane
proteins) and enhances protein digestion by providing a denaturing environment prior to
the addition of protease enzymes. ProteaseMAXTM Surfactant degrades over the course
of the digestion reaction yielding products that are compatible with downstream processes
such as mass spectrometry (MS). 10 µg of protein was used for digestion. 2 µL of 1%
ProteaseMAXTM Surfactant: 50 mM NH4HCO3 and the final reaction volume was made
up to 100 µL of reaction mixture according to the table below:
Component Volume (µL)
0.1% ProteaseMAXTM Surfactant:50 mM
NH4HCO3
20
1.0% ProteaseMAXTM Surfactant:50 mM
NH4HCO3
2
50 mM NH4HCO3 58.5
0.5M DTT 1
0.55M iodoacetamide 2.7
Trypsin (1 µg/µL) 1.8
1% ProteaseMAXTM Surfactant 1
Final Volume 100
50 mM NH4HCO3 was added to the 10 µg of protein to a final volume of 93.5 µL. 1 µL
of 0.5M DTT was additionally added and incubated at 56°C for 20 min. 2.7 µL of 0.55M
iodoacetamide was added and incubated in the dark for 15 min. 1 µL of 1%
ProteaseMAXTM Surfactant and 0.1 µg of Lys-c (Promega, #V1071) and incubated at
37°C for 6 hs. Lys-c is a protease that hydrolyses the carboxyl side of Lysine (Lys). After
this incubation, 1 µg/µL of trypsin (Promega, #V5280) was added to the mixture and
incubated at 37°C overnight. Trypsin is a protease that specifically hydrolyses arginine
and lysine residues. The condensate from the walls was collected by centrifugation at
12,000 g for 10 seconds. A final concentration of 0.5% TFA was added, mixed and
incubated for 5 min and room temperature. Sample is now ready for cleaning or peptide
concentration.
120
2.6.5 C-18 spin column peptide concentration
Peptides were purified using a C-18 spin column (Thermo Scientific, #89870) in
accordance with the manufacturer’s specifications. Buffer preparation was as follows:
Activation Solution: 50% Methanol
Equilibrium Buffer: 0.5% TFA in 5% CAN
Sample Buffer: 2% TFA in 20% CAN
Wash Solution: 0.5% TFA in 5% CAN
Elution Buffer: 70% CAN
The Column was prepared by tapping until the resin was collected at the bottom. The
column was placed into a collection tube. 200 µL of activation solution was used to rinse
the column walls and to wet the resin, centrifuged at 1500 g for 1 minute and flow-through
was discarded. This was repeated. Next, 200 µL of equilibration buffer was added and
centrifuged at 1500 g for 1 minute. This step was also repeated. The sample was loaded
into the resin bed. The column was placed into a receiver and centrifuged at 1500 g for 1
minute. Flow through was processed again following a repeat step. The column was
placed into a receiver and 200 µL of wash solution was added and centrifuged for 1
minute at 1500 g. The flow through was discarded. This wash step was repeated. The
column was placed into a new receiver tube. 20 µL of elution buffer was placed in top of
the resin and centrifuged at 1500 g for 1 minute. This step was repeated with an additional
20 µL of elution buffer to yield a final volume of 40 µL of purified peptides. Samples
were dried using a vacuum centrifuge until required.
2.6.6 Label-free mass spectrometry
Nano LC–MS/MS analysis was carried out using an Ultimate 3000 nanoLC system
(Dionex) coupled to a hybrid linear ion trap/Orbitrap mass spectrometer (LTQ Orbitrap
XL; Thermo Fisher Scientific).
A 5 μL injection of digested sample were picked up using an Ultimate 3000 nanoLC
system (Dionex) autosampler using direct injection pickup onto a 20 μL injection loop.
121
The sample was loaded onto a C18 trap column (C18 PepMap, 300 μm ID × 5 mm, 5 μm
particle size, 100 Å pore size; Dionex) and desalted for 5 min using a flow rate of 25
μL/min in loading buffer (0.1% TFA, 2% acetonitrile). The trap column was then
switched online with the analytical column (PepMap C18, 75 μm ID × 500 mm, 3 μm
particle and 100 Å pore size; (Dionex)) using a column oven at 35 °C and peptides were
eluted with the following binary gradients of:
Mobile phase buffer A (0.1% formic acid in 2% acetonitrile) and mobile phase B (0.08%
formic acid in 80% acetonitrile) 0-25% B over 280 min and 25-100% for a further 20
minuntes.
Data were acquired with Xcalibur software, version 2.0.7 (Thermo Fisher Scientific).
The Hybrid linear ion trap/Orbitrap mass spectrometer (LTQ Orbitrap XL; Thermo Fisher
Scientific) was operated in data-dependent mode and externally calibrated.
Survey MS scans were acquired in the Orbitrap in the 400–1800 m/z range with the
resolution set to a value of 30,000 at m/z 400. Up to three of the most intense ions (1+,
2+ and 3+) per scan were CID fragmented in the linear ion trap. A dynamic exclusion
was enabled with a repeat count of 1, repeat duration of 30 seconds, exclusion list size of
500 and exclusion duration of 40 seconds. The minimum signal was set to 500. All
tandem mass spectra were collected using a normalised collision energy of 35%, an
isolation window of 2 m/z, activation Q was set to 0.250 with an activation time of 30.
2.6.6.1 Identification of Proteins from Mass Spectrometry Data
The CHO fasta database was downloaded from:
ftp://ftp.ncbi.nlm.nih.gov/blast/db/FASTA consisting of 24,383 predicted genes.
2.6.6.2 Progenesis LCMS
The resulting total ion chromatograms (TIC) of quantitative label-free LC-MS data were
subjected to statistical analysis using Progenesis LC-MS software (NonLinear
Dynamics). This programme imports all runs in an “ion intensity map” format which
122
maps the ions discovered in the run, by plotting retention time against the mass/charge
ratio. All ion maps may be aligned against one reference run which allows for any shift
in retention time between runs. The intensity of each ion may then be compared against
that of all other runs, thus generating quantitative data on patterns of differential
expression between sample sets which may be analysed with statistical analysis tools such
as ANOVA, and principal component analysis.
For all experiments, unless stated otherwise, all peptides, and all proteins demonstrated
ANOVA scores of ≤0.05 for differential analysis between sample sets.
For all quantitative LC-MS/MS data, Progenesis provides the following information;
protein accession number, peptide count, number of peptides matched, confidence score,
an ANOVA value, the maximum fold change, and the average normalised abundance.
The protein accession number represents the code corresponding to the protein in NCBI-
CHO protein database. “Peptide count” refers to the number of peptides identified and
“number of peptides matched” refers to the number of peptides which were matched to
the protein. The MASCOT score (or ion score) is calculated from the combined scores of
how well each peptide matches to the protein sequence. The significance of the
differential expression between the two groups is represented by an ANOVA score (p-
value). The fold change, or average ratio, illustrates the greatest fold different in protein
ion abundance between the two groups. The average normalised abundance is calculated
from the sum of all peptide abundances from every run for each protein.
The MASCOT search parameters that were used allowed two missed cleavages, fixed
modification of cysteine (carbamidomethyl-cysteine) and variable modification of
methionine (oxidised). Peptide tolerance depended on the instrument used to acquire the
mass spectrometry data, in this case the Hybrid linear ion trap/Orbitrap mass spectrometer
used a peptide tolerance of 20 ppm. The MS/MS tolerance was set at 0.6 Da. On
completion of the database search the peptide results were filtered using MASCOT
criteria of 95% confidence interval (C.I.) threshold (p<0.05), with a minimum ion score
of ≥40. Protein identifications were accepted if they had one matched peptide.
123
2.6.7 SEAP assay
The enzymatic assay for the quantification of SEAP (secreted alkaline phosphatase) was
adapted from the method reported by Lipscomb et al., (2005). 50 µL of supernatant was
transferred to a single well on a 96 well flat bottom plate. To each sample, 50µl of 2X
SEAP reaction buffer (10.5g diethanolamine (100%), 50 µL of 1M MgCl2 and 226 mg of
L-homoarginine in a total volume of 50 mL) was added. Plates were incubated at 37 ˚C
for 10 min in order to increase enzymatic activity. After this time, 10µl of phosphates
substrate (158 mg of p-nitrophenolphosphate (SIGMA-ALDRICH®, P4744) in 5 mL of
1 X SEAP reaction buffer, made fresh for each use – adjust to suit number of sample
wells) was added to each well. A Kinetic absorbance assay was performed and the change
in absorbance per minute (OD405nm/min) was considered as an indicator of the amount of
SEAP present in the sample. Mean V was taken as the change in the average of the kinetic
slope.
2.6.8 Enzyme-linked immunosorbent assay (ELISA)
CHO-secreted IgG was quantified using the Bethyl Laboratories Human IgG ELISA
Quantitation Set (Bethyl Laboratories Inc., E80-104) in accordance with the
manufacturer’s specifications. Initially the amount of samples required including
replicate number (n = 3), blank samples and standards were calculated to determine the
number of plates required to perform an assay. 1 µL of Affinity Purified Human IgG
Coating Antibody (SIGMA-ALDRICH, A80-104A) was diluted in 100 µL of Coating
Buffer (One capsule of Carbonate-Bicarbonate Buffer (SIGMA-ALDRICH, C3041-
50CAP) in 100 mL of UHP making a 0.05M Carbonate-Bicarbonate coating buffer
solution) for each well to be coated. Add 100 µL of diluted antibody to each well and
incubate at 4˚C overnight. Wash plate 3 times using Wash Solution (SIGMA-ALDRICH,
Tris Buffered Saline, with Tween®, pH 8.0, T9039-10PAK). In order to prevent non-
specific binding, 200 µL of Blocking Solution (SIGMA-ALDRICH, Tris Buffered Saline,
with BSA, pH 8.0, T6789-10PAK) was added to each well after washing and incubated
at room temperature for 60 min. During this incubation step, standard and sample
preparation was performed. Starting from a pre-diluted (using PBS) IgG stock of 1000
ng/mL, a serial dilution was undertaken creating a range of known IgG concentrations
from 500 ng/mL-7.8 ng. A blank control of Diluent only (Blocking Solution with 0.5%
124
Tween 20) was run alongside the standards (Fig 2.11). Samples were diluted using
Diluent based on the expected IgG concentration so that the final concentration/mL fell
within the range of the standard curve.
Figure 2.11: Standard Curve Dilutions
After incubation, the blocking buffer was removed by washing 3 times as previously
described. Into each appropriate well, 100 µL of standard and sample in triplicate was
loaded and incubated at room temperature for 60 min. After, plates were washed 3 times.
Prior to addition of the HRP detection antibody (SIGMA-ALDRICH, A80-104P), a
1/50,000 dilution in Diluent was carried out by adding 1 µL of HRP Detection Antibody
to 50 mLs of Diluent. 100 µL of pre-diluted HRP detection antibody was transferred to
each sample/standard well and incubated at room temperature for 60 min. Once again,
after incubation, the HRP detection antibody was removed and the plate washed 3 times.
100 µL of the TMB substrate (SIGMA-ALDRICH, 3,3’,5,5’-Tetramethylbenzidine,
T8665-100mL) was added to each well and the enzymatic reaction was allowed to
develop at room temperature in the dark for 10-15 min yielding a colour change to a blue
solution. After 10-15 min, the reaction was stopped with the addition of 100 µL of 0.18M
(H3PO4) Phosphoric Acid (SIGMA-ALDRICH, 04102-500G) to the TMB previously
added to each well. The solution should yield a colour change from blue to yellow. Within
30 min of stopping the reaction, the enzymatic activity should be evaluated by measuring
the absorbance on an ELISA plate reader at 450 nm. Triplicate absorbance values should
be within 10% of each other. By plotting a standard curve, the elucidation of IgG content
of the unknown samples can be calculated based on the curve and taking the dilution
factor into account.
0
1
2
3
4
5
6
7
8
9
0 100 200 300 400 500 600
ABS
IgG (ng/mL)
Standard Curve IgG ng/mLStandard ng/mL
1 500
2 250
3 125
4 62.5
5 31.25
6 15.6
7 7.8
8 8
125
2.6.9 Glutamate assay
The detection of glutamate was carried out using the Glutamate Assay Kit (Sigma-
Aldrich, #MAK004) in accordance with the manufacturer’s specifications. Absorbance
readings were performed on a bench top spectrophotometer at a wavelength of 450 nm.
Glutamate standards were prepared by diluting 10 µL of the 0.1 M Glutamate standard in
990 µL of the glutamate assay buffer to make a 1 mM standard solution. From this, 0, 2,
4, 6, 8 and 10 µL into a 96 well plate forming a 0 (blank), 2, 4, 6, 8 and 10 nmole/well.
Glutamate assay buffer was added to each well to bring the volume to 50 µL. Samples
were prepared in a similar fashion topping up to 50 µL using glutamate assay buffer.
Assay reaction mixes were made up according to the scheme below:
Reagent Blank Sample Samples and Standards
Glutamate Assay Buffer 92 µL 90 µL
Glutamate Developer 8 µL 8 µL
Glutamate Enzyme Mix - 2 µL
100 µL of the appropriate Reaction Mix was added to each of the wells, mixed using a
shaker/pipetting and placed in the dark to incubate for 30 min at 37°C. The plate was
measured at 450 nm on a spectrophotometer.
2.6.10 Akta Prime IgG purification
HiTrap Protein A HP column (GE HealthCare-Life Sciences, #17-0402-01) was used for
the purification of IgG antibodies in conjunction with using the AktaPrime Plus system
(GE Healthcare). The AktaPrime Plus is a chromatography system that performs single
purifications of tagged and untagged proteins. The system is compact and comprises of a
pump, fraction collector, monitors for UV, conductivity and pH. Valves for buffer
selection, sample injection, gradient formation, and flow diversion are also integrated into
the system. The AktaPrime was cleaned using 50% ethanol and the system was purged
with PBS, including the protein A column, before each sample loading. When setting up
the AktaPrime commands, a flow rate of 1 mL/min (1 mL protein A column) with a
fraction size of mL and a pressure of 1 MPa. Load position was set to “Pos 1” for
126
equilibration and washing procedures and “Pos 8” in the case of sample loading/washing
and elution. All elutes were sent to “Waste” except in the case of sample loading, if flow-
through sample was required to be retained, and elution of IgG in which case were sent
to “load”. Finally, 0% gradient was predominantly selected except in the case of elution
in which case an 80% gradient of Buffer B (0.5% Acetic Acid) was used. After purging
the system with PBS and cleaning the protein A column using PBS (Buffer A), IgG
supernatant sample was loaded through “position 8”. The UV detector will detect all
secreted proteins as the sample flows through while IgG binds the column prior to UV
detection. After all sample has been passed though the column, Buffer A (PBS) was used
to wash the system/Column until the UV absorbance returns to baseline. An 80% gradient
was then selected for Buffer B (0.5% Acetic Acid) to flow through “position 8” and sent
to “load”. This elution was collected in waste until there was a spike in the UV detector
indicating the start of IgG elution from the column (This may take a few min – so don’t
panic). After the sample was collected, the column was washed with a 100% gradient of
Buffer B and then the system and column was washed with Buffer B in preparation for
the next sample. If finished, the system was cleaned with MS grade water loading through
both “positions 1 and 8” and sending to both “Waste” and “Load” for at least 10 column
volumes, followed with the same cleaning procedure using 20% ethanol. Once the column
was cleaned with 20% ethanol, it was also stored in 20% ethanol at 4°C. The eluted IgG
was buffer exchanged using Amicon Ultra – 4 Ultracal 10kC filters by applying the
sample to the filter and centrifuging at 4,000 g for 10 min at 4°. Flow through was
discarded and filters were topped up with 4 mL of PBS and centrifuged again following
the same conditions. This was repeated about 3-4 times and the remaining ~500 µl of
IgG-containing PBS was removed, placed in an eppi and stored at -20°C.
2.6.11 Glycomic Characterisation of Antibodies
Characterisation of antibody glycoforms was determined based on the procedure reported
by Mittermayr and Bones et al., (Mittermayr et al. 2011). Clones were cultured in a 20
mL volume spinner flask and cultured for a maximum of 4 days. Cells were removed by
centrifugation at 1500 rpm for 5 min with a centrifugation step at 4000 rpm for 10 min to
ensure removal of residual cellular debris. IgG was purified using an IgG-protein A
column on an AktaPrimeTM Plus system (see section 2.6.10 of Materials and Methods).
127
An in-solution digest was carried out using the PNGase F enzyme to remove any
asparagine N-linked glycans in addition to serine or threonine O-linked glycans. 500 µg
of total antibody protein was added to a micro-centrifuge tube in sodium bicarbonate
(SBC) making up the volume to that of the highest volume sample. As N-glycans usually
protrude from the external surface, it was not required to reduce or alkylate antibodies.
PNGase F was added to the final digest at 5% of the total volume and incubated at 37°C
overnight. A 10 kDa MWCO filter was treated with a 1:1 ethanol water solution and
centrifuged at 13,000 rpm for 5 min discarding the flow-through. Digested sample was
added to the filter and spun at 13,000 rpm for 10 min. The digestion tube was washed
with 200 µL water, added to the filter and spun at 13,000 rpm for 10 min. The filtrate was
vacuum dehydrated by centrifugation. Once dry, 20 µL of 1% formic acid was added to
the tube, vortexed briefly and reduced to dryness. This ensures complete conversion of
glycosamines present back to the reducing aldose. As monosaccharides lack
chromophores or fluorophores, it was necessary to label monosaccharaides with an
absorbing group to increase overall analytical sensitivity. To the dried N-glycans, 5 µL
of labelling solution (50 mg 2-AB, 2-aminobenzoic acid, 60 mg sodium
cyanoborohydride in 70:30 DMSO acetic acid). Mix by pipetting. An additional 5 µL of
70:30 DMSO acetic acid was added, vortexed and incubated at 65°C for two hs.
Following incubation, 90 µL of water was added to the labelling solution, mixed and
vortexed. 900 µL of acetonitrile was subsequently added, mixed and vortexed with a brief
spin to collect. Unconjugated 2-AB was removed by aspirating the liquid through a 1
mL/10 µL PhyNexus Normal Phase PhyTip preconditioned with 95% v/v acetonitrile. At
least 10 in-out cycles were performed. The tip was washed with 10 in out cycles of 95%
v/v acetonitrile venting the waste each time. The labelled N-glycans were eluted with 5 x
200 µL washes of water, collected, combined and reduced to dryness in a vacuum
centrifuge. Samples were reconstituted in an appropriate volume of water and stored at -
20°C. Ultra-Pure liquid chromatography (UPLC)-fluorescence profiling of N-glycans
was utilised to the antibody glycoform patterns in antibody samples. Hydrophilic
interaction liquid chromatography (HILIC) was used as a suitable separation technique
due to the polar nature of oligosaccharides. 2-AB labelled N-glycans were separated by
UPLC with fluorescence detection on a Waters Acquity UPLC instrument consisting of
a binary solvent manager, sample manager and fluorescence detector under the control of
Empower 2 chromatography workstation software (Water, Milford, MA). Separations
were performed using Waters BEH glycan columns, 100 x 2.1 mm i.d., 1.7 µm BEH
128
particles, using a linear gradient pf 70-53% acetonitrile at 0.56 mL/min in 16.5 min, 50
mM ammonium formate pH 4.5 was used as buffer A. An injection volume of 20 µL
sample prepared in 80% v/v acetonitrile was used throughout. Samples were maintained
at 5°C prior to injection and the separation temperature was 40°C. The fluorescence
detection excitation/emission wavelengths were λex = 330 nm and λem = 420 nm,
respectively.
2.7 Biotinylated miRNA pull-down of mRNA
The protocol utilised for the target identification of specific microRNAs including the
comparative analysis of protein complexes associated with miRNA guide and passenger
strands was based on the method published by Ørum and Lund (Orom, Lund 2007) using
Biotinylated synthetic miRNAs. Before carrying out the experimental procedure of
miRNA pull-down, 100 µL of Streptavidin-agarose beads (Sigma-Aldrich, 85881-1ML)
was incubated with 10 µL of a 10 mg/mL stock of tRNA (Sigma-Aldrich, R8508-1ML)
and 10 µL of a 10 mg/mL stock of BSA (Sigma-Aldrich, B2518-10MG) in a 1.7 mL
Eppendorf tube on a rotating platform at 4˚C for ~3 hs. This served to reduce the level of
non-specific binding. These pre-blocked beads were then washed twice in 300 µL lysis
buffer by centrifugation at 12,000 g for 30s and subsequently resuspended in 100 µL of
lysis buffer. This 100 µL of slurry was sufficient for 4 samples. To avoid separation of
bound mRNAs to our biotin-tagged miRNA as well as associated protein complexes, a
gentle lysis buffer was prepared in accordance to the recipe below.
In 100 mL of UHP water:
20 mM Tris
200 mM NaCl
2.5 mM MgCl2 (Sigma-Aldrich, M2670-100G)
0.05% Igepal (Sigma-Aldrich, 13021-50ML)
60U/mL RNase Inhibitor, Superase 20U/µL (Life-technologies, AM-2694)
1 mM DTT
Halt Protease inhibitor cocktail, EDTA-free (100X) (Thermo-fischer scientific, 87785)
The addition of the RNase inhibitor, DTT and Protease inhibitor cocktail was carried out
prior to use of the lysis buffer. For each sample, 25 µL of slurry was transferred to a 1.7
129
mL Eppendorf tube and washed twice with 500 µL of lysis buffer and resuspended in 25
µL of fresh lysis buffer. Cells were harvested 24-48 hs after transfection and washed once
with PBS by centrifugation at 1500 g for 5 min. Cells were resuspended in 1 mL lysis
buffer in a 1.7 mL Eppendorf and placed on ice. Lysed cells were then vortexed
vigorously for 10s and placed on ice for 10 min and vortexed. The lysate was cleared by
centrifugation at 12,000 g at 4˚C for 15 min in a microcentrifuge. The supernatant was
transferred to a new cold tube on ice. At this point, a 50 µL input sample was removed
from the total volume and placed on ice for further analysis. The remaining lysate was
then incubated on a rotating platform for 1 h at 4˚C with 25 µL of pre-washed
streptavidin-agarose beads. Beads incubated in lysate were collected by centrifugation for
1 minute at 5000 rpm at 4°C. The supernatant was carefully removed and the beads
washed in 500 µL of ice-cold lysis buffer. This washing step was repeated 3 times. After
the last wash, 50 µL of lysis buffer was left on the pellet. This final solution contained
Biotinylated-miRNA bound to the streptavidin-agarose beads with associated
mRNA/protein complexes ready for further processing such as RNA/Protein isolation.
2.8 Statistical analysis
A paired t-test was utilised in the case of OROBOROS measurements in miR-23 depleted
clones (Section 3.4 of results) due to the high through-put constraints within the machine
itself. Only two samples (test versus control) could be measured at any one time in
addition to a large degree of innate variance in data output. To account for this variance,
samples were statistically tested as “unequal variance” with each individual test and
control being compared individually to each other. Using excel, a paired t-test is a type 1
t-test.
A more stringent statistical method was applied for most data analysus throughout this
thesis, except in the case above. This was an unpaired t-test used to assess weather the
population means were different. This is a type 2 t-test in excel.
130
Chapter 3
Results
131
Section 3.1
Identification of microRNAs
involved in the regulatory control
of CHO cell growth rate with
subsequent phenotypic
characterisation
132
3.1.1 Identification of microRNAs involved in the regulation of CHO cell growth
rate and phenotypic characterisation
This section describes the functional assessment of a panel of miRNAs identified in a
multi-omics profiling experiment performed in our lab and described in detail in Clarke
2012 (Clarke et al. 2012).
3.1.1.1 Differential miRNA expression profile of CHO cells with varying growth
rates using an integrated miRNA, mRNA and protein expression analysis technique
Briefly, a panel of CHO clones were selected exhibiting a range of growth rates (0.011 to
0.044 hr-1) (Table 3.1) with a mean productivity of 24 (±3) pg of protein/cell/day. The
selection of sister clones derived from the same transfection pool differing only in growth
rate served to remove productivity as a variable within the profile thus enriching for
variations inherently associated with cellular proliferation.
Table 3.1: Clonal I.Ds of fast and slow CHO clones
I.D. Phenotype Specific Growth rate
X5.13F F
0.011-0.044 division/hr
X5.16F F
X5.18F F
X5.22F F
X5.2F F
X10.2S S
X11.1S S
X5.11S S
X5.1S S
X7.7S S
Of the clones represented in the table 3.1, the 10 clones were separated into two groups:
5 “fast” (≥ 0.025 hr-1) and 5 “slow” (≤ 0.023 hr-1) samples.
The incorporation of multiple “omics” disciplines served to facilitate in the identification
of non-seed miRNA targets, a problem recurrently encountered when identifying
potential targets of specific miRNAs, which is reported in greater detail in (Clarke et al.
2012). Here we focus on only the miRNAs identified within this study and pursue their
potential application in CHO cell engineering
133
3.1.1.2 Cell Culture and sampling
Clonal cell lines were grown in batch shake flask suspension culture in AS1 media
(proprietary serum-free media) at a working volume of 60 mL at 37 ˚C. Cells were
inoculated at a starting density of 2 x 105 cells/mL. Each clone was grown in triplicate
and samples were collected during mid/late exponential (log) phase (72 h). Samples were
separated into three fractions for subsequent analytical profiling of miRNA expression
using qRT-PCR based Taqman Low-Density Array (TLDA) technology, mRNA
expression using a proprietary CHO-specific WyeHamster3a oligonucleotide microarray
and protein expression using label-free LC-MS. The focus of this thesis is the miRNA
targets identified as potential engineering tools of CHO cell growth.
3.1.1.3 microRNAs associated with growth
Of the 51 miRNAs found to be DE in this study exhibiting a correlation with CHO cell
growth rate, a number have previously been associated with cellular proliferation
including miR-451 (Godlewski et al. 2010), miR-206 (Yan et al. 2009a), miR-23b* (Liu
et al. 2010a), miR-31 (Laurila et al. 2012) and miR-10a (Tan et al. 2009). Several of the
miRNAs found to be up-regulated form the miRNA cluster, miR-17 ~ 92 (miR-17, miR-
20a, miR-20b, miR-18a, miR-18b and 106a). This is a well characterised miRNA cluster
in respect to cancer (Olive, Jiang & He 2010a), having direct oncogenic transcriptional
activators such as c-myc (O'Donnell et al. 2005b) in addition to being previously
identified as highly expressed in CHO (Hackl et al. 2011).
To demonstrate the strength of this profiling strategy in reducing the false
positive/negative hits generated through utilisation of miRNA: mRNA binding criteria by
prediction algorithms (see section 1.2.6), the following is an example from the literature
in support of our findings. An experimentally validated target of miR-451 is the activator
of tumour suppression, Rab14 (Wang et al. 2011a). miR-451 was shown to be down-
regulated in fast growing CHO cells while the Rab14 protein was up-regulated in addition
to the level of expression of its mRNA remaining constant. In the case of Rab14-miR-
451, TargetScan would score Rab14 as a candidate target gene of miR-451 quite poorly
given the low degree of conservation of the 3’UTR binding site. As a result, sole
dependence on a prediction algorithm would result in the oversight of true molecular
134
targets but in cases such as this, the availability of datasets generated as described in this
current profile can aid in the prioritisation of true mRNA targets of miRNAs and
complement the output hits from prediction algorithms.
3.1.2 Phenotypic validation through transient microRNA screening
Phenotypic characterisation of a select subgroup of miRNAs derived from the list
determined to be DE in CHO cells with varying growth rates was carried out transiently
by the exogenous introduction of an artificial miRNA hairpin or duplex, termed a pre-
miR (mimic), or a single-stranded oligonucleotide fully complementary to the mature
sequence of the miRNA being targeted, termed an anti-miR/antagomiR or inhibitor (see
section 1.3.2). To ensure sufficient exogenous levels of a miRNA or adequate knockdown
of a highly expressed endogenous miRNA, 50-100 nM of mimic/inhibitor was used
during the screening process. The following section summarises the interesting results
generated from the primary phenotypic screen of the miRNAs identified from the profile
(see section 3.1.1) with the inclusion of interesting targets from literature sources/current
ongoing work.
3.1.2.1 Cell culture and transfection conditions
All transient transfections were carried out as described in section 2.3.2 of Materials &
Methods using either the transfecting reagent siSPORTTM NeoFXTM transfecting reagent
(AMBION®, AM4510) or INTERFERinTM (PolyPlus-TransfectionTM, PPLU-409-01).
Cells were seeded at a starting density of 1 x 105 cells/mL in a 50 mL aerated culti-flask
disposable bioreactors (Sartorius Stedium Biotech, DF-050MB-SSH) at a final volume of
2 mL CHO-S-SFM II (GIBCO®, 12052-098) culture media. A list of all miRNA mimics
(NBS Biologicals, M-01) and miRNA inhibitors (NBS Biologicals, M-02) used in this
transient screen are listed in the appendix (Appendix Figure A1). All appropriate
controls were included: untreated cells, transfection reagent only, a positive control
(siRNA designed against VCP) and a non-specific negative control (NC). Transfection
efficiency was based on the transfection of a siRNA designed against the mRNA coding
sequence of Vasolin-containing protein (VCP). VCP was initially identified as being
highly expressed in CHO cells with increased growth rates with subsequent siRNA-
135
mediated knockdown resulting in the induction of apoptosis and growth arrest (Doolan et
al. 2010). Inclusion of a NC (NBS Biologicals, M-03) for either a miRNA mimic or
inhibitor, served to account for the off-target effects on the cells miRNA processing
machinery. Data represented includes only the negative control and the miRNA under
investigation, however, assays yielding a poor transfection efficiency, based on siVCP,
were omitted from analysis.
Combination transfection of particular miRNAs was undertaken with optimisation carried
out as a means to determine the concentration to use that would be sufficient to potentially
mediate a phenotype without losing potency due to titrational effects. 20 nM of pre/anti-
miR was selected with a total of two miRNAs co-transfected at any one time. Each
miRNA was transfected individually (20 nM) as a means to compare any potential
synergy taking place while, however to account for the total concentration of 40 nM in
the case of co-transfected miRNAs, 20 nM of a negative control pre/anti-miR was
included for each individual miRNA (Data not shown).
After transfection, cells were cultured for 7 days at 170 rpm in a Kuhner Climo-shaker
incubator (Kuhner, Climo-shaker, ISF1-X) at 37 ˚C and assessed for growth and
productivity.
3.1.2.2 Summary of transient miRNA screen
With the focus of this thesis being that of the exploitation of miRNAs as engineering
targets to enhance the performance of CHO cells within the bioprocess, this section
summarises a subset of the list of 51 miRNAs revealed to be DE (section 3.1.1) and their
impact on the CHO cell phenotype upon transient evaluation.
From a large panel of miRNAs screened, a small subset of these miRNA was observed to
exhibit a consistent impact on CHO cell phenotype. This supports previous observations
within the literature of miRNA acting subtly, buffering cellular networks with few
miRNAs driving particular phenotypes (Mullokandov et al. 2012). The most interesting
miRNAs from the screen and their associated CHO cell phenotypes are presented in
Table 3.1.
136
Several high priority miRNAs, based on phenotypic potency, were identified such as the
tumour suppressor miR-34a, its passenger strand miR-34a* and the passenger strand of a
member of another well-known cancer-associated miRNA cluster, miR-23b*. All three
miRNAs were found to impact negatively on CHO cell viability with potent inhibition of
cellular proliferation, despite two candidates being demonstrated to be upregulated in fast
growing CHO cells. In all three instances, volumetric productivity was observed to be
reduced considerably predominantly attributed to the reduced cell numbers in early-mid
stage culture, a phenotype recurrent in instances of growth inhibition.
A large portion of the list of miRNAs screened were observed to elicit either an
inconsistent or no phenotype across replicates (data not shown). Combinatorial miRNA
transfections were explored for these inconsistent miR targets in order to exploit potential
synergy. miRNAs that exhibited mild phenotypes (in both a positive or negative
direction) or no phenotype at all were co-transfected with a certain degree of success.
Numerous miRNAs such as miR-455-3p/5p, miR-409-3p and miR-338-3p were observed
to act negatively upon transient knockdown. However, the mild and inconsistent
phenotypic shift observed for these miRNAs compared to the potent effects exhibited by
miR-34a, -34a* and -23b* over-expression resulted in these latter candidates moving to
the next level of validation as potential CHO cell engineering targets. Following
experience within our research group with miR-7 (Barron et al. 2011a), stable depletion
of miR-34a/-34a*/23b* may offer potential benefits based on their potent negative
phenotype upon their over-expression.
Despite not pursuing a small number of candidates due to resource and time constraints,
various miRNAs screened present themselves as viable candidates for stable
manipulation such as the growth inhibitor miR-206 and miR-639, the growth promoting
and potentially apoptotic protective miR-15a/-16-1 and the growth inhibitory with
possible productivity boosting attributes, miR-200a.
137
Table 3.1: Summary of all screening data in CHO-K1-SEAP
Summary of miRNA screening data in CHO
miRNA Screening Data Conclusion/Approach
miR-34a
(pre-miR)
Overexpression reduced
cellular proliferation,
induced cell death.
Knockdown improved
growth inconsistently.
Transient/stable knockdown
for the enhancement of
growth and resistance to
apoptosis
miR-23b*
(pre-miR)
Reduced cellular
proliferation, mild
induction of apoptosis
Transient/stable knockdown
for the enhancement of
growth and the resistance of
apoptosis
miR-532
(pre-miR)
Overexpression reduces
volumetric productivity.
Knockdown improves
growth rate inconsistently.
Further validation
miR-200a
(pre-miR)
Overexpression reduces
growth and improves
volumetric titre
Stable overexpression under
inducible promoter to
enhance productivity/Stable
inhibition to improve
growth
miR-181d
(pre-miR)
Inconsistent Further Validation
miR-639
(pre-miR)
Reduced cell growth Stable knockdown to
improve growth and
volumetric titre – Further
Validation
miR-455-5p/-455-3p
(anti-miR)
Inconsistent with an
indication towards
inhibitory growth effects
Further Validation
miR-206
(anti-miR)
Mild reduction in cell
growth
Further Validation
miR-34a*
(pre-miR)
Reduced cellular
proliferation, induction of
apoptosis
Transient/stable knockdown
for the enhancement of
growth and resistance to
apoptosis
miR-126 Inconsistent Further Validation
miR-338-3p
(anti-miR)
Inconsistent Further Validation
miR-378
(anti-miR)
Inconsistent Further Validation
miR-409-3p
(anti-miR)
Inconsistent Further Validation
miR-15a/16-1
(anti-miR)
Inconsistent/Increases
growth
Further Validation/Stable
knockdown
138
Section 3.2
The potential of miR-34a as an
engineering target
139
3.2 Exploring the master “tumour suppressor” miR-34a as a potential CHO
engineering target
Both miR-34a and miR-34a* were observed to reduce cell growth and induce apoptosis,
in CHO-SEAP cells upon their transient overexpression (section 3.1.2.2.2 and 3.1.2.3.1
of Results). Despite both miR-34a and miR-34a* dramatically reducing volumetric
protein output, miR-34a was observed to enhance specific productivity.
This section explores the potential of miR-34a as a candidate for CHO cell engineering
using miR-sponge technology when taking into account its conserved effect on CHO-
SEAP upon overexpression and its vast number of validated targets (Table 4.1 of
discussion).
3.2.1 Transient knockdown of miR-34a potentially enhances CHO cell growth
Given that transient overexpression of miR-34a induced an unfavourable CHO cell
phenotype it was anticipated that its transient inhibition may mediate the reverse
phenotype, i.e. accelerated growth and apoptosis resistance. However, the phenotype
observed upon its transient knockdown using anti-miRs was not as potent as hoped and
inconsistent over the course of independent assays.
A mild increase in cell density of ~13% and 30% (p ≤ 0.05) was observed on day 2 and 4
of culture, respectively (Fig. 3.1 A), with no impact on viability. Normalised productivity
was reduced by ~13% and ~25% on day 2 and 4, respectively (Fig. 3.1 B).
140
Figure 3.1: Transient miR-34a inhibition in CHO-K1-SEAP cells
Figure 3.1: Transient miR-34a inhibition was achieved using fully complementary antagonists of the
mature miR-34a commercially designed by NBS biologicals. Phenotypes under investigation were A)
Cell numbers/mL and B) Normalised productivity. Statistical significance was calculated using a
standard student t-test (p ≤ 0.05*) with an n = 3.
3.2.2 Transient inhibition of miR-34a mildly inhibits the induction of apoptosis
The transient nature of exogenously introduced anti-/pre-miRs impedes the assessment
of late stage culture phenotype such as apoptosis. Given miR-34a’s mild impact on CHO
cell viability, it was anticipated that reversing its function would positively affect viability
and prolong cell survival. We sought to induce apoptosis early in culture by mimicking
late stage culture conditions so that the transient influence of the miR-34a inhibitors has
not worn off.
0.00E+00
1.00E+06
2.00E+06
3.00E+06
4.00E+06
5.00E+06
6.00E+06
AM-SC AM-34a
Cell
Nu
mb
ers/
mL
Cell Numbers/mL
Day 2
Day 4
Day 7
0
0.00002
0.00004
0.00006
0.00008
0.0001
0.00012
AM-SC AM-34a
SE
AP
Un
its/
cell
Cell Specific Productivity
Day 2
Day 4
Day 7
A
B
*
141
3.2.2.1 Apoptosis assay development
Cell death can be induced by the supplementation of reagents such as Sodium butyrate
(NaBu) (Lee, Lee 2012) or Dimethyl sulfoxide (DMSO). However, we wanted to induce
a more realistic onset of apoptosis that would resemble the nutrient starved, hypoxic and
toxin-enriched environment of the bioreactor. This method of inducing apoptosis has been
previously reported by Druz and colleagues (Druz et al. 2011) which identified a pro-
apoptotic miRNA in CHO, miR-466h.
We grew CHO-K1-SEAP parental cells seeded at 1 x 105 cells/mL in suspension for 13
days until cell numbers and viability ceased to decline. By day 13, cultures were 15%
viable with a cell density of ~2 x 105 cells/mL (Fig. 3.2).
Figure 3.2: CHO-K1-SEAP growth curve
Figure 3.2: The growth profile for parental CHO-K1-SEAP cells as a means of determining the point
of culture at which to harvest nutrient depleted (spent) media. Cells were seeded at 1 x 105 cells/mL
and cultured until cell viability was poor (less than 20%). Cell numbers are represented in blue on
the left axis while viability is shown in red on the right axis.
As the most potent impact of a miRNA was often observed at day 4 of culture, we wanted
to establish an assay that would be monitored over the course of 4 days in an attempt to
identify the full influence of miR-34a inhibition.
We wanted to monitor the induction of cell death by culturing CHO-K1-SEAP cells
seeded at 1 x 105 cells/mL in nutrient depleted media harvested at various time points
(Fig. 3.2). Cells were cultured in spent media harvested at days 2, 3 and 4 with a fresh
media control included. It was observed that as the quality of media decreased the level
0
20
40
60
80
100
120
0.00E+00
1.00E+06
2.00E+06
3.00E+06
4.00E+06
5.00E+06
6.00E+06
0 5 10
Cell
Via
bil
ity
%
Cell
Nu
mb
ers/
mL
Day
CHO-K1-SEAP Growth Curve
Cell Numbers/mL
Viability %
142
of exponential cell growth diminished (Fig. 3.3 A). Another interesting observation was
that culture viability in all three media types declined initially but recovered (Fig. 3.3 A).
Figure 3.3: Spent medium profiling
Figure 3.3: Various levels of spent media were evaluated under batch conditions in an attempt to
elicit a mild and gradual impact on CHO cell growth and viability. Cells were seeded at 1 x 105
cells/mL and cultured in media harvested at A) days 2, 3 and 4 with a fresh media control B) day 6
and 9 with a fresh media control and C) day 9 media seeded at 4 x 105 cells/mL.
Next, day 6 and day 9 spent media was assessed. In this instance, cellular proliferation
was dramatically reduced similar to what was observed for day 4 spent media as well as
a gradual decline in cellular viability that did not recover (Fig. 3.3 B). Day 9 media
demonstrated a more defined drop in viability as far as 70% although inconsistent upon
repeat.
60
65
70
75
80
85
90
95
100
0 1 2 3 4
Via
bil
ity
%
Day
Viability Profile (Spent Media)
Day 2 Spent media
Day 3 Spent Media
Day 4 Spent Media
Fresh Media
0.00E+00
5.00E+05
1.00E+06
1.50E+06
2.00E+06
2.50E+06
0 0.5 1 1.5 2 2.5 3 3.5 4
Cell
Nu
mb
ers/m
L
Day
CHO Growth Profile (Spent Media)
60
65
70
75
80
85
90
95
100
0 0.5 1 1.5 2 2.5 3 3.5 4
Via
bil
ity
%
Day
Viability Profile (Spent Media)
60
65
70
75
80
85
90
95
100
105
0 1 2 3 4 5
Cell
ula
r V
iab
ilit
y %
Day
CHO-K1-SEAP Viability Profile
Fresh Media
Day 6
Day 9
0.00E+00
5.00E+05
1.00E+06
1.50E+06
2.00E+06
2.50E+06
3.00E+06
3.50E+06
4.00E+06
4.50E+06
5.00E+06
0 1 2 3 4 5
Cell
Nu
mb
ers/m
L
Day
CHO-K1-SEAP Growth Profile
60
65
70
75
80
85
90
95
100
105
0 1 2 3 4 5
Cell
ula
r V
iab
ilit
y %
Day
CHO-K1-SEAP Viability Profile
50
55
60
65
70
75
80
85
90
95
100
1.0E+05
1.5E+05
2.0E+05
2.5E+05
3.0E+05
3.5E+05
4.0E+05
4.5E+05
0 1 2 3 4
Cell
Via
bil
ity
%
Cell
Nu
mb
ers/m
L
Day
4 x 105 cells/mL: Day 9 SM
A
B
C
0
10
20
30
40
50
60
70
80
90
100
1.0E+05
1.5E+05
2.0E+05
2.5E+05
3.0E+05
3.5E+05
4.0E+05
4.5E+05
0 1 2 3 4
Cell
Via
bil
ity
%
Cell
Nu
mb
ers/m
L
4 x 105 cells/mL: Day 9 SM
Cell Numbers/mL
Cell Viability %
143
Finally, we attempted to increase the initial seeding density to 4 x 105 cell/mL in day 9
spent media. When this was performed there was a nice decrease in viability to a
maximum of 70% on day 4 with cell numbers declining to ~1.5 x 105 cells/mL. Upon
repetition, cells cultured in this format were not observed to revive themselves (Fig. 3.3
C).
From these results, we decided to seed at 4 x 106/mL in nutrient depleted media as it gave
a more consistent decline in both viability and cell density. However, from previous
experience, transfection at high cell density had proven troublesome. Increasing
concentrations to 4 µL of transfection reagent containing a negative control pre-miRNA
generated a similar inhibitory growth phenotype as did for siRNA-VCP (Appendix Fig.
A3). 50 nM of siRNA against VCP complexed with 4 µL of transfection reagent was the
only condition to reduce cell viability below 80% (88%) indicating a very low transfection
efficiency.
3.2.2.2 Transient inhibition of miR-34a mildly inhibits apoptosis in CHO cells
induced to undergo cell death
Several aspects of this assay were inspired by the work of Druz et al., 2011 who
performed viability assays by transfecting cells at high cell density (~106 cells) using
Lipofectamine and transferring cells into nutrient depleted media 24 h after transfection
and assayed 24 h later.
In this instance we transfected CHO cells at 1 x 105 cells/mL in 2 mL of fresh culture
media including a positive control (siRNA designed against VCP) using INTERFERinTM
(Fig. 3.4). After 48 h of growth, to allow cell density to build up, cells were pelleted and
resuspended in 2 mL of day 9 nutrient depleted media except for siRNA-VCP treated
cells in order to gauge transfection efficiency.
Transient inhibition of miR-34a in cells induced to undergo apoptosis appeared to mildly
inhibit the onset of apoptosis by 5-10% (p ≤ 0.01). Interestingly, as previously observed,
the inhibition of miR-34a also increased cellular proliferation from 1.35-2-fold (p ≤
0.001) over numerous replicates.
144
Figure 3.4: miR-34a knockdown using a modified apoptosis assay
Figure 3.4: CHO grown in spent medium were assessed for A) Cell numbers/mL and B) Viability 24
h subsequent to exposure to spent media, n = 3, t-test were calculated using a standard student t-test.
3.2.3 Stable depletion of miR-34 using a microRNA sponge
With the previous observation that transient inhibition of miR-34a improved CHO cell
growth and mildly inhibited apoptosis, we thought it encouraging to explore its stable
depletion in both SEAP and IgG-secreting CHO cells. miR-34a has been observed to be
expressed at quite low levels in CHO cells (Hackl et al. 2011) using next-generation
sequencing while its counterparts miR-34b/c were expressed quite abundantly. With the
single nucleotide difference lying outside the seed region, both miR-34b and miR-34c are
members of the same seed family as miR-34a thus potentially targeting the same mRNAs
0.00E+00
5.00E+04
1.00E+05
1.50E+05
2.00E+05
2.50E+05
3.00E+05
3.50E+05
VCP AM-NC AM-34a
Cell
Nu
mb
ers/
mL
Cell Numbers/mL
24hr
0
10
20
30
40
50
60
70
80
90
100
VCP AM-NC AM-34a
Via
bil
ity
%
Viability %
24hr
A
B
***
**
0
10
20
30
40
50
60
70
80
90
100
VCP AM-NC AM-34a
Via
bil
ity
%
Viability %
24hr
145
(Bader 2012). Stable depletion of miR-34a using a fully complementary binding sequence
downstream of a deGFP sponge reporter will effectively target all three miRNAs, miR-
34a/b/c. Hence forth, miR-34a when in reference to its depletion through a sponge decoy
will be referred to as miR-34.
3.2.3.1 miR-34 sponge design and generation
The miR-34 miR-sponge sequence was designed to be fully complementary to its specific
target miRNA with the exception of base-pair mismatches or “wobble” introduced at
nucleotide locations 10-12 and placed downstream of a deGFP reporter (Fig. 3.5). This
wobble was introduced to reduce miRNA-RISC-mediate mRNA cleavage that has been
characterised to occur at nt 9-10 thus prolonging the functional half-life of the sponge
construct (Kluiver et al. 2012a).
146
Figure 3.5: miR-34 sponge construct and miR interaction
Figure 3.5: The construct that forms the miR-34 sponge, A) a destabilised GFP reporter under the
transcriptional activation of a CMV promoter and a downstream polyA tail. Sandwiched between
the deGFP coding sequence and polyA tail is the multiple cloning site (MCS) containing XhoI/EcoRI
used for insertion of the miR-34 sponge sequence. This sequence contains four concatemeric binding
sites for the miR-34 family in tandem with 5 nucleotide spacers and forms an artificial 3’UTR of
deGFP. Wobble is introduced into the sponge sequence across nt 10-12 (highlighted in red) which
serves to prolong the life time of the sponge mRNA as mRNA cleavage should be inhibited. An
additional selection gene is incorporated into the sponge construct, Hygromycin resistance (HYGRO
or HYG), which is used for the generation of stable CHO populations. B) The mature sequence of
miR-34a with the seed sequence highlighted in blue.
The sequence of the miR-34 sponge construct including restriction sites is shown in table
3.2.
Table 3.2: miRNA sponge sequence design for miR-34
The miR-34 sponge sequence pEX-A-miR-34 (Fig. 3.6 A) was amplified in DH5α cells
to achieve a working concentration suitable for molecular cloning (see material and
5’-TCTAGACTCGAG...........acgcgACAACCAGCTCCAACACTGCCAacgcg...........GAATTCTCTAGA-3’
EcoRI
3’-TGTTGGTCGA TGTGACGGT-5’TTC
XhoI EcoRI
XbaI XbaI
A
B
5’- TGGCAGTGTCTTAGCTGGTTGT- 3’
TTC
CTT
5’-3’ microRNA Sponge sequence
miR
Sponge
XbaI
Res.
XhoI
Res.
Sponge seq. EcoRI
Res.
XbaI
Res.
miR-
34a
TCTA
GA
CTCG
AG
acgcgACAACCAGCTCCAACACTGCCAac
gcgACAACCAGCTCCAACACTGCCAacga
gACAACCAGCTCCAACACTGCCAtgacgt
ACAACCAGCTCCAACACTGCCAtcatc
GAA
TTC
TCTA
GA
147
methods section 2.4.17 for transformation protocol). Removal of the sponge sequence
was carried out using the restriction enzymes XhoI/EcoRI. Linearization of the pdeGFP
plasmid backbone was carried out using the same enzymes to generate “sticky-ends”.
Both miR-34 sponge insert and backbone were isolated by gel extraction (Fig. 3.7 A and
B, see Materials and methods section 2.4.14). After ligation of the miR-34 sponge
sequence into the pd2EGFP-sponge vector, several colonies were picked upon ampicillin
selection and mini-prepped to isolate plasmid DNA. Two restriction sites NotI and XbaI
were selected to diagnose successful miR-34 sponge insertion (Fig. 3.7 C). A positive
colony (1 or 4 from Fig. 3.7 C) was re-amplified to a working stock and diagnosed using
the same enzymes (Fig. 3.7 D). miR-34 sponge vector was sequenced using the following
primer:
EGFP-701-For 5’-GGACGAGCTGTACAAGAAGC-3’
Sequence alignment can be seen in appendix figure A4.
The miR-34 sponge and control sponge plasmid was transfected into SEAP and 1.14
secreting cells (see Materials and Methods section 2.3.3). 24 h after transfection, cells
were selected in the presence of 350 µg/mL Hygromycin for 3 passages generating a
mixed pool (MP) population. Stable cells were adapted to suspension culture in the
presence of PVA (anti-clumping agent) before characterisation.
148
Figure 3.6: Plasmid maps for pEX-A (carrying sponge sequences) and microRNA Sponge backbone
Figure 3.6: Plasmid maps for pEX-A (carrying miR-sponge sequences) and microRNA sponge backbone: A) Bacterial expression vector harbouring the
microRNA sponge concatemers with an ampicillin resistance marker (proprietary plasmid from MWG Eurofins) B) microRNA sponge backbone containing
multiple cloning site (MCS) with cloning restriction sites, GFP reporter construct and Hygromycin (HYGR) resistance marker.
< Sponge Sequence>
↓
A B
Xba I (2375)
Not I(1795)
149
Figure 3.7: Diagnostic gels for miR-34 sponge vector preparation
Figure 3.7: Diagnostic gels for miR-sponge preparation A) digests of the sponge backbone and a pre-
made sponge negative control using restriction enzymes XhoI/EcoRI B) miR-34 sponge insert
extracted using XhoI/EcoRI ran on a 2% agarose gel to confirm is successful extraction C) mini-prep
digestion of miR-34 sponge bacterial colonies tested using restriction enzymes XbaI/NotI liberating
successful inserts of ~750 bp. Control sponge digestions using the same restriction enzymes were
carried out to include no-insert control (NIC), sponge-NC (negative control), raw sponge backbone
(BB) with no insert and miR-7 sponge.
100bp
200bp
A B
C
100bp
500bp
850bp
miR-34a Sponge
500bp
850bp
A
B
D
150
3.2.3.2 Confirmation of miR-34 sponge binding efficacy
To assess the binding efficiency of the miR-34 sponge for endogenous miR-34 family
members, we transiently introduced mature miR-34a pre-miRNA into both miR-34 and
miR-NC sponge mixed pool populations. The varying GFP intensities indicated the
generation of mixed populations for both plasmids (Random integration) (Fig. 3.8 A).
Cells expressing the miR-34 sponge exhibited a lower average GFP expression when
assessed flow cytometry when compared to miR-NC spg (Fig. 3.8 B). This reduced
average GFP expression was potentially due to endogenous miR-34 interacting with and
repressing GFP translation. Transient introduction of a pre-miRNA duplex specific for
miR-34a resulted in a 90% (p ≤ 0.001) reduction in GFP fluorescence in the case of the
miR-34 sponge mixed pool (Fig. 3.8 B). A reduction in GFP expression was also observed
in the case of miR-NC sponge upon miR-34a overexpression but not as dramatic as in the
case of miR-34 sponge, possible a side-effect of miR-34’s known impact on cell growth.
The average reduced GFP expression in miR-34 sponge cells was reflected again when a
clonal panel was isolated from both mixed pools potentially due to endogenous miR-34
(Fig. 3.9). A range in GFP expression can also be observed for all sponge panels due to
random integration. The GFP intensity was observed to be much higher in CHO-1.14
cells stably transfected with miR-34 and miR-NC sponge constructs. This could
potentially be due to a reduced transfection efficiency in the case of CHO-SEAP cells.
Furthermore, the integration efficiency or transcriptional activity may be higher in the
case of CHO-1.14s contributing to this significant GFP expression difference.
151
Figure 3.8: miR-34 sponge binding efficiency
Figure 3.8: The binding efficacy of the miR-34 sponge construct when miR-34a was transiently over
expressed. A) Visually demonstrates the intensity of GFP in both mixed populations (miR-34 and
miR-NC sponge) transfected with either pre-miR-NC or pre-miR-34a. B) GFP intensity of both
mixed pools is further demonstrated using the Guava ExpressPlus programme. Statistical analysis
was carried out using a standard student t-test (p ≤ 0.001***).
pre-miR-NC pre-miR-34a
miR-34a spg
miR-NC spg
0
500
1000
1500
2000
2500
pre-miR-NC pre-miR-34a
Av
e G
FP
ex
press
ion
miR-34a spg vector binding efficacy
miR-NC spg MP
miR-34a spg MP
***
A
B
152
Figure 3.9: GFP profile of miR-34 sponge clonal panels in CHO-SEAP and CHO-
1.14 cells
Figure 3.9: The GFP expression of 24 clones isolated from both miR-34 and miR-NC sponge mixed
populations using FACS sorting in both CHO-K1-SEAP and CHO-1.14 cells. Average GFP
expression calculated across each panel condition is represented as a broken red line.
3.2.3.3 miR-34a overexpression potentially inhibits conserved targets
Prior to miR-34 sponge characterisation, we sought to identify potential mRNA targets
that miR-34a was regulating. A small list of validated miR-34a targets were screened
upon transient overexpression of miR-34a in CHO-K1-SEAP cells. Targets were
identified and chosen from the review by Andreas G. Bader (Bader 2012) (Table 4.1 in
Discussion). Primers were designed based on the CHO sequence genome database
(http://www.chogenome.org/). Primer sequences are shown in table 2.1 of materials and
methods section 2.4.3. RNA was harvested 24 h after transfection (see section 2.4.6 of
Materials and Methods) and reverse transcribed (see section 2.4.2 of Materials and
Methods). Semi-quantitative PCR was carried out on a small subset of targets including
E2F3, BCL2, CCND1, DCR-1, CDK4, CDK6 and FUT8 using β-actin as an endogenous
control (materials and methods section 2.4.3).
0
100
200
300
400
500
600
700
800
7 11 22 17 13 4 20 21 19 12 24 23 9 10 2 6 3 14 8 1 16 15 5 18
GF
P E
xp
ress
ion
Cone
miR-34a sponge clonesCHO-K1-SEAP
GFP
0
100
200
300
400
500
600
700
800
20 11 8 19 14 21 16 23 1 9 12 5 2 3 7 6 10 4 13 18 22 15 24 17
GF
P E
xp
ress
ion
Clone
miR-NC sponge clonesCHO-K1-SEAP
GFP
Ave GFP = 436.7 Ave GFP = 301.11
0
500
1000
1500
2000
2500
3000
5 2 6 1 22 7 21 4 17 20 23 11 14 16 10 3 8 13 18 19 15 12 24 9
GF
P E
xp
ress
ion
Clone
miR-NC sponge clonesCHO-1.14
GFP
0
500
1000
1500
2000
2500
3000
9 8 12 1 24 18 3 15 7 4 2 10 5 23 21 16 6 11 20 13 19 14 22 17
GF
P E
xp
ress
ion
miR-34a sponge clonesCHO-1.14
GFP
0
500
1000
1500
2000
2500
3000
5 2 6 1 22 7 21 4 17 20 23 11 14 16 10 3 8 13 18 19 15 12 24 9
GF
P E
xp
ress
ion
Clone
miR-NC sponge clonesCHO-1.14
GFP
0
500
1000
1500
2000
2500
3000
5 2 6 1 22 7 21 4 17 20 23 11 14 16 10 3 8 13 18 19 15 12 24 9
GF
P E
xp
ress
ion
Clone
miR-NC sponge clonesCHO-1.14
GFP
Ave GFP = 965.85 Ave GFP = 796.13
CHO-K1-SEAP
CHO-1.14
miR-NC sponge panel miR-34a sponge panel
153
Cells were assayed 48 h after transfection as a means to confirm both transfection
efficiency and retention of the miR-34a induced phenotype (Fig. 3.10 A). Gel-based
visualisation indicated that several cell cycle-related targets were under the influence of
miR-34a with CCND1 potentially the most dramatic (Fig. 3.10 B).
Real-time quantitative (qRT)-PCR using SYBR® Green (see section 2.4.4 of Materials
and Methods) showed a 50% (p ≤ 0.001) reduction in the mRNA abundance of CCND1
(Fig. 3.10 C).
Figure 3.10: semi-quantitative and qRT-PCR identification of miR-34a targets
Figure 3.10: Assessment of validated miR-34a targets upon miR-34a over-expression (PM-34a)
compared to a negative control (PM-NC) in addition to a positive control (siRNA-VCP) A) Influence
of miR-34a overexpression on proliferation and viability in CHO-SEAP cells 24 h post transfection.
B) Gel-based semi-quantitative PCR using a panel of primers for validated targets. C) Real-time
quantitative PCR on CCND1 upon miR-34a overexpression when compared to pre-miR-NC
transfection. Statistical analysis was calculated based on cyclic threshold (Ct) values using a standard
student t-test (p ≤ 0.001***).
E2F3
BCL2
CCND1
DCR-1
CDK4
CDK6
FUT8
0.00E+00
5.00E+04
1.00E+05
1.50E+05
2.00E+05
2.50E+05
3.00E+05
VCP PM-NC miR-34a
Cel
l N
um
ber
s/m
L
Cell Numbers/mL @ 48 hours
miR-34a Overexpression
40
50
60
70
80
90
100
VCP PM-NC miR-34a
Via
bil
oil
ty %
Cell Viability % @ 48 hours
miR-34a Overexpression***
***
0
0.2
0.4
0.6
0.8
1
1.2
1.4
pre-miR-NC pre-miR-34a
RQ
CCND1
CCND1***
0
0.2
0.4
0.6
0.8
1
1.2
1.4
pre-miR-NC pre-miR-34a
RQ
CCND1
CCND1***
A
B
C
siRNA-
VCP
siRNA-
VCP PM-NC PM-34a
PM-NC PM-34a
PM-NC PM-34a
E2F3
BCL2
CCND1
DCR-1
CDK4
CDK6
FUT8
Via
bil
ity %
Cel
l N
um
ber
s/m
L
154
3.2.3.4 miR-34 sponge influence on endogenous miR-34a and CCND1
To further explore the influence of the miR-34 sponge on the endogenous levels of miR-
34a, we evaluated the levels of miR-34a in a stable CHO-1.14 clone expressing our miR-
sponge. We further assessed the levels of CCND1, which were shown to be reduced by
50% upon transient miR-34a overexpression, in the same clone to identify if its levels
were increased.
miR-34a abundance was assessed using Singleplex Taqman qRT-PCR (section 2.4.8 of
Materials and Methods) on three miR-NC-spg clones and three miR-34 spg clones. A
~5-fold reduction in the mature levels of miR-34a (p ≤ 0.001) was observed in clones
engineered with the miR-34 sponge when compared to non-specific controls (Fig. 3.11
A).
Next siRNA-mediated knockdown of the miR-34 sponge construct was carried out in
miR-34 sponge clone 21.
Knockdown of the miR-34 sponge was achieved using a siRNA designed against the GFP
(siGFP) gene sequence. RNA was harvested 72 h after transfection and qRT-PCR was
carried out for mature miR-34a expression. As observed before, the miR-34 sponge clone
exhibited a 5-7-fold reduction in mature miR-34a levels when compared to the miR-NC
sponge control (Fig. 3.11 B). However, a 2-fold increase in the mature levels of miR-34a
was observed in the case of miR-34 sponge treated with siGFP compared to miR-NC-
treated miR-34 sponge (Fig. 3.11 B). Knockdown of the miR-34 sponge was
demonstrated by a 90% (p ≤ 0.001) reduction in GFP expression (Fig. 3.11 B inset). The
levels of miR-34a however did not return to levels comparable with that of the miR-NC
sponge control (Fig. 3.11 B) which could potentially be due to the presence of residual
sponge construct. This does however, indicate that the miR-34 sponge construct is
specific for miR-34.
Finally, we attempted to assess the impact that the miR-34 sponge construct exerted on a
previously validated target of miR-34a, CCND1. The greatest difference was observed in
the case of miR-34 sponge clone 21 which demonstrated a 1.2-3.5-fold increase in
endogenous CCND1 levels when compared to all three control sponge clones (Fig. 3.11
C). However, the abundance of CCND1 in 2 of the 3 miR-34 sponge clones (3 and 4)
exhibited lower levels when compared to 2 of the 3 control clones (15 and 20).
155
Figure 3.11: The influence of stable miR-34 sponge expression on miR-34a and
CCND1 abundance
Figure 3.11: Impact of the miR-34 specific sponge on the endogenous levels of mature miR-34a and
the target CCND1. A) Taqman qRT-PCR for miR-34a in a panel of miR-34 and control sponge
clones. B) [Inset] The knockdown of the miR-34 sponge construct using siGFP compared to pre-miR-
NC control. Endogenous levels of miR-34a in a miR-34 sponge clone untreated (Cells only) and
transfected with a control (pre-miR-NC) and siRNA for GFP (siGFP). Endogenous levels of miR-34a
are also shown for an untreated control sponge clone (miR-NC spg 2) compared to cells only miR-34
spg 21 C) qRT-PCR for the validated miR-34a target, CCND1.
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
miR-NC spg 2
miR-NC spg 15
miR-NC spg 20
miR-34a spg 3
miR-34a spg 4
miR-34a spg 21
RQ
miR-34a expression
miR-34a exp
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
RQ
miR-34a expression
*** *** ***
0
100
200
300
400
500
600
- pre-miR-NC deEGFP siRNA
GF
P e
xp
ress
ion
Treatment
miR-34a spg-21CHO-1.14
GFP
***
0
1
2
3
4
5
6
7
8
9
no condition pre-miR-NC siRNA-deGFP no condition
RQ
miR-34a exp
miR-34a exp
0
1
2
3
4
5
6
7
8
9
no condition pre-miR-NC siRNA-deGFP no condition
RQ
**
miR-34 spg 21 miR-NC spg 2
A
B
C
0
1
2
3
4
5
6
miR-NC spg 2
miR-NC spg 15
miR-NC spg 20
miR-34a spg 3
miR- 34a spg 4
miR-34a-spg 21
RQ
CCND1 mRNA miR-34a sponge
CCND1 exp
0
1
2
3
4
5
6
miR-NC spg 2
miR-NC spg 15
miR-NC spg 20
miR-34a spg 3
miR- 34a spg 4
miR-34a-spg 21
RQ
CCND1 mRNA miR-34a sponge
Cells only Cells onlypre-miR-NC siGFP
Cells only pre-miR-NC siGFP
2 15 20 3 4 21miR-34 spgmiR-NC spg
2 15 20 3 4 21miR-34 spgmiR-NC spg
156
3.2.4 Evaluation of miR-34 depletion in CHO mixed pool populations
Two CHO cell lines (CHO-SEAP and CHO-1.14) secreting recombinant human-secreted
alkaline phosphatase (SEAP) and the more industrially relevant antibody (IgG) were
generated using the miR-34 sponge construct discussed earlier in section 3.2.3.1. Both
stable mixed pools were evaluated under several culture conditions in an attempt to
identify the potential benefit of the stable inhibition of miR-34.
3.2.4.1 miR-34 depletion in CHO-K1-SEAP cells
miR-34 sponge CHO-SEAP mixed pools - Batch
Stable miR-34 sponge mixed populations were cultured in batch mode over a 6 day period
and sampled every 48 h along with a non-specific control sponge population. Cells were
seeded initially at 2 x 105 cells/mL in a 5 mL culture volume of CHO-S-SFM II in the
absence of Hygromycin selection reagent.
There was a 55% (p ≤ 0.05) improvement in cell density on day 2 of culture (Fig. 3.12
A). Repeated experiments demonstrated increased growth over various stages of the
culture process. Furthermore, cellular viability appeared to be mildly enhanced by 6-15%
(p ≤ 0.05) late in culture across all three replicates (Fig. 3.12 B).
No impact was observed on volumetric SEAP yield when both test and controls were
compared (Fig. 3.12 C). However, the increase in cell numbers on day 2 of culture
correlated with a reduction in normalised productivity (Fig. 3.12 D).
157
Figure 3.12: Stable miR-34 sponge mixed pool CHO-SEAP under batch conditions
Figure 3.12: The influence of stable miR-34 depletion in batch mode is demonstrated compared to a
negative control sponge. A) Cells numbers sampled on day 2,4 and 6 B) Cellular viability C)
Volumetric SEAP yield/mL designated as SEAP Units D) Normalised SEAP activity.
miR-34 sponge CHO-SEAP mixed pool – Fed-batch
Feeding strategies are commonly implemented in biopharma as a means to further boost
maximal cell densities and mediate a prolonged viable production phase. An initial feed
of 15% (v/v) of CD CHO-Efficient Feed A was supplemented on day 0 of culture with
subsequent feeding of 10% (v/v) every 3 days of culture. Cell growth was dramatically
enhanced in the case of the non-specific control by almost 2-fold at its apex upon feed
addition, however, this was not observed in the case of miR-34 pools (Fig. 3.13 A).
Cellular viability on day 8 of culture was reduced in the miR-34 sponge population by
~20% (Fig. 3.13 B). However, both volumetric SEAP and specific productivies were
observed to be enhanced by a range of about 1.6 to 3-fold in miR-34 depleted pools (Fig.
3.13 C and D).
0
5
10
15
20
25
0 1 2 3 4 5 6 7
No
rm
ali
sed
Pro
du
cti
vit
y (
x1
0-5
)
Day
0
0.5
1
1.5
2
2.5
3
3.5
4
4.5
5
0 1 2 3 4 5 6 7
Cell
Nu
mb
ers/
mL
(x1
06)
Day
0.00E+00
5.00E+05
1.00E+06
1.50E+06
2.00E+06
2.50E+06
3.00E+06
3.50E+06
4.00E+06
4.50E+06
5.00E+06
0 2 4 6 8
Cell
Nu
mb
ers/
mL
miR-34a spg MP
miR-NC spg MP
miR-34a spg MP
0
20
40
60
80
100
120
0 1 2 3 4 5 6 7
Via
bil
ity
%
Day
miR-34a spg MP
0
50
100
150
200
250
300
350
400
0 1 2 3 4 5 6 7
SE
AP
Un
its/m
L
Day
miR-34a sponge MP
A B
DC
*
*
158
Figure 3.13: Stable miR-34 sponge mixed pool CHO-SEAP under fed-batch
conditions
Figure 3.13: Fed-batch (feed represented as a broken orange line) results for the miR-34 sponge
mixed pools alongside a non-specific control sponge assessed over the course of a 6 day culture. A)
Cell Numbers/mL B) Cellular viability C) Volumetric SEAP yield D) Specific SEAP productivity (n
= 3), statistics were carried out using a paired student t-test (p ≤ 0.001***).
3.2.4.2 miR-34 depletion in IgG-secreting CHO-1.14 cells
miR-34 sponge CHO-1.14 mixed population - Batch
A 40% (p ≤ 0.01) improvement in cell density was recorded on day 2 of culture with a
20% (p ≤ 0.001) increase observed on day 4 and 6 when compared to controls (Fig. 3.14
A). Cell density was improved consistently across separate replicates but varied in
magnitude with viability appearing unaffected (Fig. 3.14 B). Volumetric IgG (see section
2.6.8 of Materials and Methods) was determined to be unchanged with a small increase
of about 10% late in culture (Fig. 3.14 C). This could potentially be due to the elevated
cell numbers, as cell specific productivity was reduced (Fig. 3.14 D).
0
5
10
15
20
25
30
35
0 2 4 6 8 10
No
rma
lise
d p
rod
uct
ivit
y (
x1
0-5
)
Day
0
2
4
6
8
10
12
0 2 4 6 8 10
Cel
l N
um
ber
s/m
L (
x1
06)
Day
0
10
20
30
40
50
60
70
80
90
100
0 2 4 6 8 10
Via
bil
ity
%
Day
miR-34a spg viabiltiy FB
0
50
100
150
200
250
300
350
400
0 2 4 6 8 10
SE
AP
Un
its/
mL
Day
miR-34a spgCHO-K1-SEAP MP Fed-Batch
0
10
20
30
40
50
60
70
80
90
100
0 2 4 6 8 10
Via
bil
ity
%
Day
miR-34a spg viabiltiy FB
miR-NC spgMP
miR-34a spgMP
***
A B
C D
159
Figure 3.14: Stable miR-34 sponge mixed pool CHO-1.14 under batch conditions
Figure 3.14: CHO-1.14 cells stably engineered to inhibit miR-34 were cultured under batch
conditions and sampled over the course of a 6 day batch process reporting on the phenotypes, A) Cell
numbers/mL B) Cellular viability C) Volumetric titres, IgG ng/mL and D) Cell specific productivity.
3.2.5 Generation of miR-34 sponge clones
As a means of further investigating the phenotype observed within the mixed populations,
we isolated a panel of clones from the mixed pools using fluorescent activated cell sorting
(FACS, see section 2.3.4 of Materials and Methods).
3.2.5.1 GFP fluorescence across clonal panels
Clones were selected that exhibited a medium-to-high GFP intensity as a means to isolate
clones that expressed the miR-34 sponge at levels sufficient to sequester all endogenous
miR-34a/b/c. A range of GFP intensities were observed across all clonal panels in both
cell lines (Fig. 3.9). Both miR-34 sponge clone panels were assessed for performance in
1 mL batch suspension culture.
0
0.5
1
1.5
2
2.5
0 1 2 3 4 5 6 7
Cel
l N
um
ber
s/m
L (
x1
06)
Day
**
***
***
0
20
40
60
80
100
120
0 1 2 3 4 5 6 7
Ce
llula
r V
iab
ility
%
Day
miR-34a spg MPA B
C D
0
10
20
30
40
50
60
70
80
90
100
0 2 4 6 8 10
Via
bil
ity
%
Day
miR-34a spg viabiltiy FB
miR-NC spgMP
miR-34a spgMP
0
0.5
1
1.5
2
2.5
3
3.5
4
4.5
0 1 2 3 4 5 6 7
IgG
ng
/mL
(x1
05)
Day
0
20
40
60
80
100
120
140
160
180
0 1 2 3 4 5 6 7
Sp
ecif
ic P
rod
uct
ivit
y (
pg
/cel
l/d
ay
)
Day
160
3.2.5.2 Impact of miR-34 depletion across a CHO-SEAP cell panel
Previous characterisation in mixed pools indicated that in fed-batch culture, stable miR-
34 inhibition potentially enhanced specific productivity while batch conditions mildly
enhanced CHO cell growth with no benefit to volumetric SEAP yield. Upon clonal
isolation, clones were seeded at 1 x 105 cells/mL and cultured over a 6 day suspension
batch process in parafilm sealed 24-well culture plates. As in the case of previous batch
processes, samples were taken every 48 h and assessed for the primary phenotypes,
growth, viability and productivity. Phenotypic characterisation was carried out across the
entire panel of clones, both test and control, in an effort to assess the average impact of
miR-34 depletion in addition to isolating exceptional clones.
When cultured in a 24-well format, miR-34 depleted clones did not display a similar
phenotype to what was indicated in mixed pool batch culture. Growth was reduced by
~30% (p ≤ 0.01), ~75% (p ≤ 0.001) and ~80% (p ≤ 0.01) on day 2, 4 and 6, respectively
(Fig. 3.15). Furthermore, although the average performance was reduced, not a single
miR-34 depleted clone demonstrated higher growth than the top control performers. This
reduced growth was observed in 1 mL culture despite mixed pools and subsequently
selected clones demonstrating “normal” growth patterns in 5 mL.
161
Figure 3.15: CHO-SEAP miR-34 sponge clonal panel – Cell Numbers/mL
Figure 3.15: Growth of individual miR-34 and miR-NC stable sponge clones in 1 mL batch
conditions. Average panel performance at each time point is represented as a broken red line. Clones
highlighted in an orange box are clones that were subsequently selected for further characterisation.
Statistical analysis was carried out using a standard student t-test (p ≤ 0.01** and p ≤ 0.001***).
0.00E+00
2.00E+05
4.00E+05
6.00E+05
8.00E+05
1.00E+06
1.20E+06
1.40E+06
1.60E+06
24 21 9 18 11 10 1 8 20 22
Cell
Nu
mb
ers/
mL
Clone
miR-34a spgCHO-K1-SEAP Cells/mL
Day 2
0.00E+00
1.00E+06
2.00E+06
3.00E+06
4.00E+06
5.00E+06
6.00E+06
7.00E+06
6 15 17 24 10 22 2 9 19 13 8 1 21 11 20
Via
ble
Cell
s/m
L
Clone
miR-NC spg Cells/mL Day 4
Cells/mL
0.00E+00
1.00E+06
2.00E+06
3.00E+06
4.00E+06
5.00E+06
6.00E+06
9 24 17 6 10 8 21 13 11 19 22 15 1 20 2
Via
ble
Cell
s/m
L
Clone
miR-NC spg Cells/mL Day 6
Cells/mL
0.00E+00
2.00E+05
4.00E+05
6.00E+05
8.00E+05
1.00E+06
1.20E+06
1.40E+06
1 17 15 10 6 22 9 24 2 13 11 21 8 19 20
Via
ble
Cell
s/m
L
Clone
miR-NC spg Cells/mL Day 2
Cells/mL
0.00E+00
1.00E+06
2.00E+06
3.00E+06
4.00E+06
5.00E+06
6.00E+06
7.00E+06
21 9 1 18 8 24 10 20 11 22
Cell
Nu
mb
ers/
mL
Clone
miR-34a spg panel Day 4
0.00E+00
1.00E+06
2.00E+06
3.00E+06
4.00E+06
5.00E+06
6.00E+06
1 20 10 8 11 18 24 21 22 9
Cel
l Nu
mb
ers/
mL
Clone
miR-34a spg Day 6
Day 6
5.31 x 105 Cells/mL7.63 x 105 Cells/mL
8.03 x 105 Cells/mL
4.98 x 106 Cells/mL
3.55 x 106 Cells/mL
1.22 x 106 Cells/mL
miR-NC spg panel miR-34a spg panel
Clone
0.68-fold
0.24-fold
**
***
**
0.22-fold
Cel
l N
um
ber
s/m
L (
x1
05)
Cel
l N
um
ber
s/m
L (
x1
06)
Cel
l N
um
ber
s/m
L (
x1
06)
1
0
2
3
4
5
6
0
1
2
3
4
5
6
7
0
2
4
6
8
10
12
14
Day 2
Day 4
Day 6
162
Figure 3.16: CHO-SEAP miR-34 sponge clonal panel – Cellular Viability
Figure 3.16: Clonal viability in both miR-34 and miR-NC sponge stable CHO-SEAP panels on day 6
of culture. There was no gross difference in clonal cell viability on the preceding time points of day 2
and 4. Average viability across each clonal panel is represented as a broken red line. Clones selected
for scale up are highlighted in an orange box. Statistical analysis was carried out using a standard
student t-test (p ≤ 0.001***).
The majority of miR-34 depleted clones preformed significantly worse than the panel of
miR-NC spg clones with a considerable reduction in average viability from 56% in the
case of the control panel to 29% (p ≤ 0.001) in the case of the miR-34 test panel (Fig.
3.16).
0
10
20
30
40
50
60
70
80
90
100
1 20 22 8 10 18 24 11 21 9
Cell
ula
r V
iab
ilit
y %
Clone
miR-34a day 6 viability %
0
10
20
30
40
50
60
70
80
90
100
11 9 8 24 17 10 20 6 22 19 1 13 21 15 2
Via
bil
ity
%
Clone
miR-NC spg Viability (Day 6)
miR-NC spg panel miR-34a spg panel
Ave Via 5% = 56.1% Ave Via % = 29%
***
163
Figure 3.17: CHO-SEAP miR-34 sponge clonal panel – Volumetric SEAP yield/mL
Figure 3.17: Volumetric output of SEAP/mL of each clone in both the miR-34 and miR-NC sponge
panels. The average SEAP yield across each panel is represented as a broken red line while each
clone subsequently selected for scale up is highlighted in an orange box. Statistical analysis was
carried out using a standard student t-test (p ≤ 0.01** and p ≤ 0.001***).
A significant 3.64-fold (p ≤ 0.001), 2.13-fold (p ≤ 0.01) and 1.97-fold (p ≤ 0.001) increase
was observed when the average volumetric SEAP yields of the miR-34 sponge panel was
compared to the miR-NC spg panel on days 2, 4 and 6, respectively (Fig. 3.17).
0
100
200
300
400
500
600
700
800
9 18 8 1 11 10 24 20 22 21
SE
AP
Un
its/
mL
Clone
miR-34a spongeCHO-K1-SEAP panel (Day 4)
0
100
200
300
400
500
600
700
800
11 2 21 22 1 8 20 15 6 19 9 13 10 17 24
SE
AP
Un
its/
mL
Clone
miR-NC spongeCHO-K1-SEAP panel (Day 4)-50
0
50
100
150
200
250
300
11 20 22 21 15 8 2 6 1 19 12 10 9 17 24
SE
AP
Un
its/
mL
Clone
miR-NC SEAP Day2
SEAP/mL
0
50
100
150
200
250
300
9 11 24 18 8 10 1 20 22 21
SE
AP
Un
its/m
L
Clone
miR-34a spongeCHO-K1-SEAP panel
Day 2
0
100
200
300
400
500
600
700
800
11 8 2 20 21 15 1 22 6 13 19 9 10 24 17
SE
AP
Un
its/
mL
Clone
miR-NC sponge SEAP/mL Day 6
0
100
200
300
400
500
600
700
800
24 18 9 1 10 8 11 20 22 21
SE
AP
Un
its/
mL
Clone
miR-34a spg panel Day 6
Day 6
Day 2
Day 4
Day 6
Ave SEAP = 36.15
Ave SEAP = 150.76
Ave SEAP = 221.27 Ave SEAP = 429.35
Ave SEAP = 322.3
Ave SEAP = 124.12
miR-NC spg panel miR-34a spg panel
***
***
**
3.64-fold
2.13-fold
1.97-fold
164
Figure 3.18: CHO-SEAP miR-34 sponge clonal panel – Specific SEAP productivity
Figure 3.18: Normalised SEAP productivity per cell for each clone within both miR-34 and miR-NC
sponge panels. Average normalised productivity is represented as a broken red line while clones
selected for scale up are highlighted in orange boxes. Normalised productivity was calculated based
on the value of SEAP produced divided by the number of cells present. The outlier (miR-34 spg clone
9-day 6 has a value of 7.96 x 10-3. Statistical analysis was calculated using a standard student t-test (p
≤ 0.05* and p ≤ 0.001***).
An increase of 3.83-fold (p ≤ 0.001), 8.14-fold (p ≤ 0.001) and 14.9-fold (p ≤ 0.05) was
observed in normalised productivity for the miR-34 sponge panel (Fig. 3.18).
There was no indication that viability was enhanced by the stable depletion of miR-34.
Furthermore, at this scale, growth appeared to be diminished in comparison to control
clones. For these reasons, to eliminate bias, clones were selected based on their
0.00E+00
5.00E-04
1.00E-03
1.50E-03
2.00E-03
2.50E-03
3.00E-03
9 24 22 20 11 21 8 10 18 1
SE
AP
Un
its/
Cel
l
Clone
miR-34a spg Day 6 Qp
0.00E+00
1.00E-04
2.00E-04
3.00E-04
4.00E-04
5.00E-04
22 11 8 9 10 18 1 24 20 21
SE
AP
Un
its/
Cel
l
Clone
miR-34a spgCHO-K1-SEAP panel Qp
Day 2
0.00E+00
1.00E-04
2.00E-04
3.00E-04
4.00E-04
5.00E-04
20 11 21 8 22 2 15 19 6 13 1 9 10
SE
AP
Un
its/
Cel
l
Clone
miR-NC spgCHO-K1-SEAP panel Qp
Day 2
0.00E+00
1.00E-04
2.00E-04
3.00E-04
4.00E-04
5.00E-04
6.00E-04
7.00E-04
8.00E-04
20 11 21 2 1 8 22 19 15 13 6 9 10 17 24
SE
AP
Un
its/
Cell
Clone
miR-NC spongeCHO-K1-SEAP panel, Qp (Day 4)
0.00E+00
1.00E-04
2.00E-04
3.00E-04
4.00E-04
5.00E-04
6.00E-04
7.00E-04
8.00E-04
11 10 20 22 9 18 8 24 1 21S
EA
P U
nit
s/ce
llClone
miR-34a spongeCHO-K1-SEAP panel, Qp (Day 4)
miR-NC spg panel miR-34a spg panel
Ave SEAP /Cell = 6.74 x 10-5 Ave SEAP /Cell = 2.5 x 10-4
Ave SEAP /Cell = 4.4 x 10-5 Ave SEAP /Cell = 3.59 x 10-4
Ave SEAP /Cell = 1.5 x 10-3Ave SEAP /Cell = 1.0 x 10-5
3.83-fold
8.14-fold
14.9-fold
***
***
*
0.00E+00
5.00E-04
1.00E-03
1.50E-03
2.00E-03
2.50E-03
3.00E-03
2 20 1 11 15 8 21 22 6 19 13 10 9 24 17
Cel
l N
um
ber
s/m
L
Clone
miR-NC spg Qp Day 6
SE
AP
Un
its/
Cel
l (x
10
-4)
SE
AP
Un
its/
Cel
l (x
10
-4)
SE
AP
Un
its/
Cel
l (x
10
-3)
0
0.5
1.5
2.5
1
2
3
0
1
2
3
4
5
6
7
8
0
1
2
3
4
5Day 2
Day 4
Day 6
165
volumetric productivity attributes, a process that was indicated to be enhanced during the
fed-batch cultivation in mixed pools. Clones 8, 10 and 18 were selected from the miR-34
sponge panel while clones 2, 11 and 20 were selected from the miR-NC sponge panel
(highlighted in orange boxes, Fig. 3.17). The clones selected demonstrated a spread
within the other phenotypic categories such as viability and growth; however, attempting
to encompass all performance categories into the one clone would have compromised the
selection of high SEAP producers and would ultimately give an inappropriate indication
of performance. The clones selected were progressed for further study in 5 mL culture
scale up. By selecting clones based on their enhanced productivity attributes, this could
potentially introduce a bias when clones are scaled up to higher culture volumes. Clones
have been seen across the literature to perform differently within various culture formats.
By selecting clones based on phenotypic attributes other than those suggested to be
enhanced during mixed population assays, a bias can be introduced thereby selecting
clones whose observed phenotypes are potentially not a result of miRNA intervention.
This should be considered throughout this thesis for all clones.
3.2.5.3 Impact of miR-34 depletion across a CHO-1.14 cell panel
We previously showed that the CHO-1.14 mixed pools grew to better cell densities in
batch culture without any impact on cell viability or product yield. This improvement was
not maintained when pool-derived clones were cultured in 24-well plates.
Cell growth across all time points was reduced by 22% (p ≤ 0.05), 8% and 39% (p ≤ 0.01)
on days 2, 4 and 6, respectively when compared to non-specific controls (Fig. 3.19).
Additionally, no clone within the miR-34 sponge panel demonstrated a superior growth
capacity to any of the top performing controls.
166
Figure 3.19: CHO-1.14 (IgG) miR-34 sponge clonal panel – Cell Numbers/mL
Figure 3.19: The individual growth performance of the clones that make up the miR-34 and miR-NC
sponge panels with the average growth being represented as a broken red line. Clones that were
subsequently selected for scale up are highlighted in orange boxes. Statistical analysis was carried
out using a standard student t-test to include all clones (p ≤ 0.05* and p ≤ 0.01**).
Cellular viability remained constant in the early stages of culture (Days 2 and 4, Data not
included). Within each population there were some poor performers but the average
viability across miR-NC and miR-34 sponge was 63.4% and 66%, respectively (Fig.
3.20).
0.00E+00
2.00E+05
4.00E+05
6.00E+05
8.00E+05
1.00E+06
1.20E+06
20 16 9 3 21 4 13 1 6 17 14 2 11 15 8 22
Cell
Nu
mb
ers/
mL
Clone
miR-NC sponge panel CHO-1.14
0.00E+00
2.00E+05
4.00E+05
6.00E+05
8.00E+05
1.00E+06
1.20E+06
15 4 24 13 11 5 16 21 12 10 9 3 8 6 22 14 17 19 23 20
Cell
Nu
mb
ers/
mL
Clone
miR-34a spg CHO-1.14 Day 2
0.00E+00
5.00E+05
1.00E+06
1.50E+06
2.00E+06
2.50E+06
3.00E+06
3.50E+06
4.00E+06
20 21 16 9 3 15 13 6 14 17 4 1 11 2 8 22
Cell
Nu
mb
ers/
mL
Clone
miR-NC sponge panel CHO-1.14 Day 4
0.00E+00
5.00E+05
1.00E+06
1.50E+06
2.00E+06
2.50E+06
3.00E+06
3.50E+06
4.00E+06
21 16 15 10 22 5 13 11 3 24 19 6 4 17 14 8 9 23 12 20
Cell
Nu
mb
ers/
mL
Clone
miR-34a spg CHO-1.14 Day 4
0.00E+00
5.00E+05
1.00E+06
1.50E+06
2.00E+06
2.50E+06
3.00E+06
3.50E+06
4.00E+06
4.50E+06
16 9 15 3 6 14 4 21 13 2 8 20 1 11 17 22
Cell
Nu
mb
ers/
mL
Clone
miR-NC sponge panel CHO-1.14 Day 6
0.00E+00
5.00E+05
1.00E+06
1.50E+06
2.00E+06
2.50E+06
3.00E+06
3.50E+06
4.00E+06
4.50E+06
10 21 11 13 4 5 3 6 9 24 16 22 8 15 12 14 23 20 19 17
Cell
Nu
mb
ers/
mL
Clone
miR-34a spg CHO-1.14 Day 6
miR-NC spg panel miR-34a spg panel
6.08 x 105 Cells/mL 4.78 x 105 Cells/mL
1.97 x 106 Cells/mL 1.83 x 106 Cells/mL
1.26 x 106 Cells/mL2.05 x 106 Cells/mL
*
**
0.78-fold
0.92-fold
0.61-fold
Cel
l N
um
ber
s/m
L (
x1
05)
Cel
l N
um
ber
s/m
L (
x1
06)
Cel
l N
um
ber
s/m
L (
x1
06)
0
2
4
6
8
10
12
4
3.5
2.5
2
1.5
1
0.5
0
3
4.5
4
3.5
3
2.5
2
1.5
1
0.5
0
Day 2
Day 4
Day 6
167
Figure 3.20: CHO-1.14 (IgG) miR-34 sponge clonal panel – Cell Viability %
Figure 3.20: The cellular viability of each individual clone of both miR-34 and miR-NC stable sponge
clonal panels with the average viability across each panel being represented as a broken red line.
Clones that were selected for scale up are highlighted in orange boxes.
When volumetric IgG levels were assessed, there were a large number of clones that did
not produce IgG above the detectable limits of the ELISA assay. Although these poor
performing clones did pull down the average performance across the population, there
were still some reasonably high performing clones within both groups. There was an
average reduction of 23%, 36% and 28% observed in volumetric IgG on days 2, 4 and 6
of culture, respectively (Fig. 3.21). No single clone from the miR-34 sponge panel
demonstrated a volumetric IgG output above the top performing clone from the control
group. However, clones that did produce IgG did so above the average level of the control
panel.
0
10
20
30
40
50
60
70
80
90
100
3 13 2 9 15 16 6 1 11 4 8 14 17 21 20 22
Via
bil
ity
%
Clone
miR-NC sponge CHO-1.14 Day 6 Via
0
10
20
30
40
50
60
70
80
90
100
9 6 4 8 3 10 21 5 11 16 24 13 23 12 17 20 22 19 15 14
Via
bil
ity
%
Clone
miR-34a spg Day 6 Via%
miR-NC spg panel miR-34a spg panel
Ave Via = 63.4% Ave Via = 66%
168
Figure 3.21: CHO-1.14 (IgG) miR-34 sponge clonal panel – Volumetric productivity
Figure 3.21: Volumetric IgG secreted product of all individual clones within both the miR-34 and
miR-NC sponge clonal panels. The average IgG product across each panel is represented as a broken
red line. Clones subsequently selected for scale up are highlighted in an orange box.
0.00E+00
1.00E+04
2.00E+04
3.00E+04
4.00E+04
5.00E+04
6.00E+04
7.00E+04
8.00E+04
9.00E+04
22 17 16 6 4 8 3 10 21 9 19 20 14 5 23 13 11 24 15 12
IgG
ng
/mL
Clone
miR-34a IgG Vol Day 2
0.00E+00
1.00E+04
2.00E+04
3.00E+04
4.00E+04
5.00E+04
6.00E+04
7.00E+04
8.00E+04
9.00E+04
2 14 8 4 20 3 6 11 15 21 13 1 17 16 22 9
IgG
ng
/mL
Clone
miR-NC sponge CHO-1.14 IgG/mL Day 2
0.00E+00
5.00E+04
1.00E+05
1.50E+05
2.00E+05
2.50E+05
3.00E+05
3.50E+05
20 2 14 21 15 8 4 3 6 13 11 22 17 16 9 1
IgG
ng
/mL
Clone
miR-NC sponge CHO-1.14 IgG/mL Day 60.00E+00
5.00E+04
1.00E+05
1.50E+05
2.00E+05
2.50E+05
14 20 2 8 3 4 15 6 21 11 13 17 9 1 16 22
IgG
ng
/mL
Clone
miR-NC sponge CHO-1.14 IgG/mL Day 4
0.00E+00
5.00E+04
1.00E+05
1.50E+05
2.00E+05
2.50E+05
21 4 6 16 3 17 22 10 8 9 19 14 20 5 15 24 23 11 13 12
IgG
ng
/mL
Clone
miR-34a spg IgG Day 4
0.00E+00
5.00E+04
1.00E+05
1.50E+05
2.00E+05
2.50E+05
3.00E+05
3.50E+05
3 4 22 10 6 16 21 8 17 9 14 19 15 23 20 12 5 13 11 24
IgG
ng
/mL
Clone
miR-34a spg IgG Vol Day 6
miR-NC spg panel miR-34a spg panel
Ave = 3.4 x 104 ng/mL
Ave = 8.57 x 104 ng/mL
Ave = 1.72 x 105 ng/mL Ave = 1.25 x 105 ng/mL
Ave = 6.67 x 104 ng/mL
Ave = 2.64 x 104 ng/mL
0.77-fold
0.64-fold
0.72-fold
IgG
ng
/mL
(x
10
5)
IgG
ng
/mL
(x
10
5)
IgG
ng
/mL
(x
10
4)
0
1
1.5
2
2.5
3
3.5
4
4.5
5
0
0.5
1
1.5
2
2.5
0
0.5
1
1.5
2
2.5
3
3.5
Day 4
Day 2
Day 6
169
It is not surprising that when the specific productivities of each panel was calculated the
values observed reflected that of the proliferative capacity of each clone versus their
volumetric output (Fig. 3.22).
Figure 3.22: CHO-1.14 (IgG) miR-34a sponge clonal panel – specific productivity
Figure 3.22: Specific IgG productivity of each individual clone from both the miR-34 and miR-NC
sponge panel. The average specific productivity is represented as a broken red line. Clones selected
for scale up are highlighted in an orange box.
The performance of the clonal panel did not reflect the phenotype determined within the
predecessor mixed pool population. From this perspective we decided to select those
clones that displayed good volumetric IgG characteristics. Clones selected are highlighted
in orange boxes in all the graphs. The clones selected from the miR-34 sponge panel
0
20
40
60
80
100
120
17 8 22 3 6 4 16 10 21 9 20 14 23 19 12 15 5 24 13 11
IgG
pg
/cel
l/d
ay
0
10
20
30
40
50
60
70
80
90
17 21 6 3 4 8 22 16 10 9 20 14 23 19 12 5 11 24 13 15
IgG
pg
/cel
l/d
ay
0
20
40
60
80
100
120
140
160
22 17 6 16 8 4 3 10 21 9 20 19 23 14 5 11 13 24 12 15
IgG
pg
/cel
l/d
ay
0
20
40
60
80
100
120
140
160
14 8 2 4 6 11 15 20 3 21 13 22 1 17 16 9
IgG
pg
/cel
l/d
ay
0
10
20
30
40
50
60
70
80
90
14 2 8 4 15 6 20 3 11 21 13 22 17 1 9 16
IgG
pg
/cel
l/d
ay
Day 2
Day 4
0
20
40
60
80
100
120
22 2 8 4 14 20 15 21 6 11 3 13 17 1 9 16
IgG
pg
/cel
l/d
ay
Day 6
miR-34 spg panelmiR-NC spg panel
Clone
Ave Qp = 59.65
pg/cell/day
Ave Qp = 46.46
pg/cell/day
Ave Qp = 25.39
pg/cell/day
Ave Qp = 21.49
pg/cell/day
Ave Qp = 31.88
pg/cell/day
Ave Qp = 31.57
pg/cell/day
0.77-fold
0.80-fold
0.67-fold
*
**
170
were clones 3, 4 and 21. These clones displayed reasonable phenotypes across all
categories with an emphasis on volumetric productivity. Additionally, miR-NC sponge
clones 2, 15 and 20 were selected based on similar criteria.
3.2.6 Phenotypic evaluation of miR-34 depleted clones in a 5 mL volume
miR-34 depletion has been observed to elicit two distinct phenotypes in two recombinant
CHO cell lines, SEAP and IgG. In CHO-SEAP mixed populations, volumetric SEAP
yield was enhanced in fed-batch while clones in 1 mL batch displayed enhanced SEAP
yield. CHO-1.14 mixed pools demonstrated increased growth with reduced specific
productivity while clone panels showed reduced performance across all traits (Growth,
yield and specific productivity). This section explores the potential of miR-34 depletion
in individual clones selected from both the CHO-SEAP and CHO-1.14 panels isolated for
their production characteristics.
3.2.6.1 Selected miR-34 depleted CHO-SEAP clones
miR-34 sponge CHO-SEAP clones - Batch
All miR-34 depleted clones were observed to achieve a reduced maximal cell density of
between 20-50% when compared to the two control clones (miR-NC spg 2 and 11) that
reached the highest cell density (Fig. 3.23 A). However, when compared to the worst
performing control clone, all miR-34 sponge clones demonstrated 1.9-2.3-fold higher cell
numbers (Fig. 3.23 A). This control clone would be considered an outlier as the “normal”
growth profile of the CHO-SEAP cell line would be greater than this (~3.5 x 106 cell/mL
in the case of miR-NC mixed pools). All miR-34 sponge clones demonstrated reduced
viability on day 6 of culture (Fig. 3.23 B) with two clone in particular (miR-34 spg clone
9 and 18) exhibiting the lowest viability at ~1%. Furthermore, these miR-34 sponge
clones appeared to clump early in culture (days 3-4) contributing to a sudden drop in
apparent cell density.
171
Figure 3.23: miR-34 sponge clones culture under batch conditions
Figure 3.23: Batch culture results in a 5 mL volume for miR-34 depleted CHO-SEAP clones selected
from the clonal panel when compared to non-specific control clones. Phenotypes assessed were A)
Cell growth, B) Culture viability, C) SEAP yield/mL and D) Normalised productivity. Statistical
analysis was calculated using a standard student t-test (p ≤ 0.05* and an n = 3).
All miR-34 sponge clones displayed an enhanced SEAP secretion when compared to the
two lowest producing control clones (Fig. 3.23 C). One control sponge clone (Clone 11)
did reach comparable maximal SEAP levels when compared to miR-34 sponge clones
(Fig. 3.23 C). Two miR-34 sponge clones (9 and 10) demonstrated a high normalised
productivity potentially due to their propensity to clump and display low cell viability
resulting in a low cell density on day 6 of culture (Fig. 3.23 D). miR-34 sponge clones
exhibited a tendency to clump early in culture regardless of the addition of PVA or the
process of cloning out non-clumping cells from these cultures. At this point, Fed-batch
was not pursued and assessment of miR-34 depletion in CHO-SEAP cells was terminated.
3.2.6.2 Selected miR-34 depleted CHO-1.14 clones
miR-34 sponge CHO-1.14 clones - Batch
3 clones were selected from both control and test panels based on their volumetric
productivity. However, a growth benefit was observed in the case of mixed pools that was
Cel
l N
um
ber
s/m
L (
x1
06)
Day
SE
AP
Un
its/
mL
Day
Via
bil
ity
%
Day
-2.00E-03
0.00E+00
2.00E-03
4.00E-03
6.00E-03
8.00E-03
1.00E-02
1.20E-02
1 2 3 4 5 6 7
SE
AP
Unit
s/ce
ll
Day
No
rma
lise
d p
rod
uct
ivit
y (
x1
0-4
)
Day
A B
C D
*
172
not retained in 1 mL batch culture. When scaled up to a 5 mL volume, all clones
demonstrated enhanced growth of 1.3-5-fold (p ≤ 0.001) on day 2 and 1.5-3-fold (p ≤ 0.01
and 0.001) on day 4 (Fig. 3.24 A). While all three clones reached higher cell numbers on
day 6 when compared to two control clones (miR-NC spg 2 and 15), one control clone
had increased cell numbers as well as comparable cell viability (Fig. 3.24 A and B).
Despite the varied performance of the control clones, siRNA-mediated knockdown of the
miR-34 sponge construct in miR-34 sponge 21 caused a 30% (p ≤ 0.001) reduction in
density on day 3, suggesting that the presence of the sponge had a growth benefit (Fig.
3.25). Efficient knockdown of the miR-34 sponge was previously demonstrated (Fig. 3.11
B).
All three miR-34 sponge clones demonstrated enhanced viability on day 6 of culture of
between 65-80% while the miR-NC sponge control clones exhibited viability between
21-53% (Fig. 3.24 B). However, one control clone (miR-NC spg clone 20) demonstrated
a viability comparable to miR-34 sponge clones. This would suggest that miR-34
depletion is conferring a potential benefit by delaying the onset of apoptosis and that miR-
NC clone 20 is displaying high viability due to the nature of clonality, selecting clones
with inherent beneficial attributes, by chance.
When volumetric IgG productivity and specific productivity were assessed, there was a
dramatic reduction in the production capacity of all three miR-34 sponge clones of around
10-60% across all time points when compared to the range of titres produced by all three
control clones (Fig. 3.24 C). As cell numbers were observed to be enhanced in the case
of the miR-34 depleted clones, specific productivity was further reduced to a range of 10-
80% (Fig. 3.24 D).
173
Figure 3.24: CHO-1.14 stable miR-34 sponge clones under batch growth conditions
Figure 3.24: Clones selected from each panel (miR-34 and miR-NC sponge in batch conditions in a
scaled up 5 mL culture volume. A) Cell numbers/mL B) Cellular viability C) Volumetric IgG
secretion and D) Specific productivity. Statistical analysis was determined using a standard student
t-test with each miR-34a sponge clone being compared to all three controls. (p ≤ 0.01** and p ≤
0.001***).
Figure 3.25: sponge knockdown mediates a reversal in the growth phenotype
Figure 3.25: The impact that miR-34 sponge knockdown exerted on the growth profile of a stable
miR-34a sponge clone (clone 21) upon transient transfection of a siRNA designed against the deGFP
(siGFP) reporter construct. Cells were sampled on day 3 of culture after transfection with GFP
fluorescence inhibition reported in Figure 3.11.
0
0.5
1
1.5
2
2.5
3
3.5
0 1 2 3 4 5 6 7
Cel
l N
um
ber
s/m
L (
x1
06)
Day
0.00E+00
5.00E+05
1.00E+06
1.50E+06
2.00E+06
2.50E+06
3.00E+06
3.50E+06
0 2 4 6 8
Cell
Nu
mb
ers/
mL
Day
miR-34a spg clones batch
miR-NC spg 2
miR-NC spg 15
miR-NC spg 20
miR-34a spg 3
miR-34a spg 4
miR-34a spg 21
0
20
40
60
80
100
120
0 1 2 3 4 5 6 7
Cell
ula
r V
iab
ilit
y %
Day
miR-34a spg clones batchA B
DC
***
**
***
0
0.2
0.4
0.6
0.8
1
1.2
1.4
0 1 2 3 4 5 6 7
IgG
ng
/mL
(x1
05)
Day
0.00
10.00
20.00
30.00
40.00
50.00
60.00
0 1 2 3 4 5 6 7
Sp
ecif
ic p
rod
uct
ivit
y (
pg
/cel
l/d
ay
)
Day
0
10
20
30
40
50
60
70
80
90
100
0.00E+00
2.00E+05
4.00E+05
6.00E+05
8.00E+05
1.00E+06
1.20E+06
1.40E+06
1.60E+06
- pre-miR-NC deEGFP siRNA
Cell
s/m
L
Treatment
miR-34a spg-21CHO-1.14
Cells/mL
Viability %
***
Cells only pre-miR-NC siGFP
1.6
1.4
1.2
1.0
0.8
0.6
0.4
0.2
0
Cel
l N
um
ber
s/m
L (
x1
06)
174
miR-34 sponge CHO-1.14 clones – Fed-batch
Clone 3 from the miR-34 sponge panel was omitted from this Fed-batch study as it
exhibited the poorest growth phenotype. Once again, miR-34 sponge clones achieved an
increased maximal cell density of 1.5 to 2-fold when compared to two of the three control
clones (Fig. 3.26 A). Furthermore, both miR-34 depleted clones demonstrated an
extended productivity phase by maintaining viability above 80% until day 8 (Fig. 3.26
B). However, as observed in the case of batch culture, one control sponge exhibited a
comparable phenotype, miR-NC spg clone 20, when both growth and viability were
assessed.
A reduction of about 20-40% was observed for both miR-34 depleted clones when
compared to control clones 15 and 20 while a large reduction of around 80% was observed
in the case of the high producing clone miR-NC spg 2 in IgG productivity (Fig. 3.26 C).
When specific productivity was determined, there was a 40-90% reduction in both miR-
34 sponge clones when compared to all three control clones (Fig. 3.26 D).
Figure 3.26: CHO-1.14 stable miR-34 sponge clones under fed-batch growth
conditions
Figure 3.26: Fed-batch culture for the selected clones engineered to stably deplete miR-34 and the
non-specific control. Addition of feed media is represented as a broken orange line. Phenotypes
reported are A) Cell numbers/mL B) Cell viability C) Volumetric IgG secretion and D) specific
productivity.
0
10
20
30
40
50
60
70
80
90
100
0 2 4 6 8 10
IgG
ng
/mL
Day
0.00E+00
5.00E+05
1.00E+06
1.50E+06
2.00E+06
2.50E+06
3.00E+06
0 2 4 6 8 10
Cell
Nu
mb
ers/
mL
Day
miR-34a spg clonesCHO-1.14
0
20
40
60
80
100
120
0 2 4 6 8 10
Via
bil
ity
%
Day
miR-34a spg clonesCHO-1.14
miR-NC-2
miR-NC-15
miR-NC-20
miR-34a-4
miR-34a-21
0
20
40
60
80
100
120
0 2 4 6 8 10
Via
bil
ity
%
Day
miR-34a spg clonesCHO-1.14
0.00E+00
5.00E+04
1.00E+05
1.50E+05
2.00E+05
2.50E+05
0 2 4 6 8 10
IgG
ng
/mL
Day
miR-34a spgCHO-1.14
Cel
l N
um
ber
s/m
L (
x1
06)
0
0.5
1
1.5
2
2.5
0
0.5
1
1.5
2
2.5
3
IgG
ng
/mL
(x
10
5)
A B
C D
IgG
pg
/cel
l
175
From this section it can be seen that the depletion of miR-34 mediated improved growth
in CHO-1.14 cells. However, this growth benefit did not push the cells to reach higher
cell densities than previously observed for the parental population with the addition of
potentially compromising the production capacity of each clone. Although miR-34
sponge clones did perform better in regards to growth and viability when compared to 2
of the 3 control clones selected, one control clone exhibited comparable phenotypic
attributes.
With miR-34 proving an interesting miRNA but potentially not a viable miRNA for CHO
engineering, we decided to explore its potential role in influencing the glyocoprofile of
secreted IgG.
3.2.7 miRNA-mediated interference of CHO cell antibody glycoforms
Section 1.4 introduced the importance of antibody glycoforms to the functionality and
potency of antibody-mediated effector functions such as ADCC. Increased ADCC has
shown to be enhanced for main-stream anti-cancer therapeutics such as Herceptin through
the manipulation of the core fucose added to the N-acetylglucosamine at the conserved
Asparagine-297 (Suzuki et al. 2007a)(Li et al. 2013b) and Rituxan (Li et al. 2013b). The
report that miR-34a could interact with the 3’UTR of FUT8 (Bernardi et al. 2013a) and
interfere with its mature protein levels led us to explore the stable depletion of miR-34
contributing to an altered antibody glycoform, fucosylation, in the case of recombinant
CHO-1.14 cells. The protocol followed is as described by Mittermayr and Bones et al.,
(Mittermayr et al. 2011) and section 2.6.11 in Materials and Methods.
Two clones were selected from the CHO-1.14 parental population engineered to stably
deplete miR-34 including a separate non-specific sponge clone. When the antibody
glycoform profile was determined for both the selected miR-34 sponge and miR-NC
sponge clone, it was observed that there was no change in the peak ratios between each
sample indicating no alteration in the fucosylation profile (G0F, G1F and G2F (see
section 1.4.2 of Introduction for nomenclature, Fig. 3.27).
176
Figure 3.27: Glyco-profile of a miR-34 sponge clone versus miR-NC sponge clone
Figure 3.27: The peak intensities and elution times (min) for the N-glycans isolated and fluorescently
labelled with 2-AB (2-aminobenzoic acid) with the core fucosylated N-acetylglucosamine of interest
being G0F, G1F and G2F, i.e. biantennary N-glycan no, one or two galactose groups on either
terminal N-acetylglucosamine. GU indicates glucose units
However, one interesting observation was that the intensity of the peak profile was
considerably higher in the case of the secreted antibody derived from the miR-34 sponge
clone when compared to the miR-NC sponge clone (Fig. 3.27).
As a means to ensure that antibody quantification was accurate, potentially contributing
to the difference in peak intensities, 10 µg of each antibody was run on a 4-12% Bis-Tris
SDS gel and stained with Coomassie Blue. One antibody pair was reduced using
iodoacetamide to break apart disulphide bonds, liberating the individual heavy and light
chains.
Coomassie blue stained gel showed that there was similar quantities of IgG antibody in
both miR-NC and miR-34 sponge clone samples (Fig. 3.28).
6
miR-34a sponge
miR-NC sponge
G0F
G1F
G2FG0
G1
G2F+S
G2F+2S
GO = Biantennary with no galactose.GO = Biantennary with no galactose, no core fucose.G1, G2 = Plus 1 or 2 galactose residues, respectively.F = Addition of a fucose.
7 854GU 9 10 11
177
Figure 3.28: Coomassie blue semi-quantification of secreted IgG from miR-34 and
miR-NC sponge clones
Figure 3.28: 10 µg of purified non-reduced and reduced IgG antibody secreted from a miR-NC and
miR-34 sponge CHO-1.14 clone stained with Coomassie Blue.
Following these observations, we designed several treatments of the miR-34 sponge clone
21 as a means to assess the stable depletion of miR-34 on influencing the IgG
glyocoprofile.
Firstly, as the intensity of the glyocoprofile peaks were higher in the case of the miR-34
sponge compared to the control, we assessed the glyocoprofile of the parental CHO-1.14
from which both sponge clones were derived. It was anticipated that the parental CHO-
1.14 would demonstrate a similar low intensity profile to that of the negative control.
When this was assessed, both parental and miR-34 sponge clones exhibited a similar
glyocoprofile of equal intensity (Fig. 3.29 Sample A compared to Sample C or E).
Secondly, if miR-34 depletion was mediating this increased peak intensity, we attempted
to knockdown the sponge using a siRNA against GFP with the hope of observing a
reduced intensity when compared to the control miR-34 sponge treated with pre-miR-
NC. Successful inhibition of the sponge was demonstrated previously for this clone (Fig.
250
100
55
35
25
NC-IgG 34a-IgG NC-IgG 34a-IgG
15
10
Non-reduced Reduced
178
3.11 B inset). However, no change in peak intensity was observed (Fig. 3.29 Sample D
compared to Sample C or E).
From the perspective of fucosylation, CHO cells secret antibody of which >90% are
fucosylated (Jefferis 2009a). It is possible that by depleting miR-34 and increasing the
expression of FUT8, an increase in fucosylation would not be apparent. We next
transiently over-expressed miR-34a in a stable miR-34 sponge clone with the anticipation
of observing a shift in the fucosylation profile towards reduced fucosylation. No shift in
the three peaks associated with fucosylation (G0, G1 and G2) was observed (Fig. 3.29
Sample B compared to Sample C or E).
179
Figure 3.29: Glyco-profile of IgG produced from a miR-34 sponge clone 21 treated
with pre-miR-34a and siRNA for GFP
Figure 3.29: The IgG glyco-profiles of various sample conditions including A) CHO-1.14 parentals
untransfected, B) miR-34a sponge clone 21 transfected with pre-miR-34a, C) miR-34a sponge clone
21 transfected with a pre-miR-NC control, D) miR-34a sponge clone 21 transfected with a siRNA
specific for the GFP in the miR-34a sponge construct and E) miR-34a sponge clone 21 transfected
with a pre-miR-NC control. The primary fucosylated glycan peaks are labelled including G0F, G1F
and G2F. GU indicates glucose units.
The focus of CHO cell line engineering towards glyco-manipulation has proven to be a
promising endeavour from literature reports. Further evaluation of the impact of miR-34a
on the fucosylation capacity in CHO cells would be required starting by assessing the
mature protein levels of the FUT8 gene upon overexpression of miR-34a. Additionally,
several IgG producing cell lines would have to be assessed for their resulting glycoforms
following miR-34a transfection. Increased miR-34a expression elicits an undesirable
phenotype such as cell death and growth inhibition. From this perspective, the
identification of alternative miRNA targets of FUT8 that do not contribute this unwanted
phenotype would be invaluable.
CHO 1.14 Parentals
miRNA 34a SPG 21 (34a Treated)
miRNA 34a SPG 21 (NC Treated)
miRNA 34a SPG 21 (siRNA Treated)
miRNA 34a SPG 21 (NC Treated)
Sample A
Sample B
Sample C
Sample D
Sample E
GU 4 5 6 7 8 9 10 11
180
Section 3.3
Stable depletion of independent
microRNA duplex components:
miR-23b and miR-23b*
181
3.3. CHO cell engineering using miR-23b and miR-23b*
Several miRNAs that were screened, such as miR-23b*, impacted negatively on the CHO
cell phenotype despite their positive correlation with improved growth in the profiling
experiment. This section describes and characterises two miRNA partners, miR-23b and
its passenger miR-23b*, selected for stable CHO cell line engineering within both a SEAP
and IgG-secreting cell line derived from the same CHO-K1 parental background.
3.3.1. Selection of miR-23b and miR-23b* for stable knockdown
Transient miR-23b* overexpression inhibited growth by 90% and cell viability was
reduced by 30%. However, it was anticipated that stable inhibition of this miRNA might
induce a reversed phenotype, enhancing growth and potentially inhibiting cell death. Past
experience in CHO using miR-7, demonstrated that overexpression of miR-7 caused
growth inhibition (Barron et al. 2011a) while stable inhibition improved CHO cell growth
and product yield (Sanchez et al. 2013b).
The mature guide strand miR-23b was selected based on previous identification and
evaluation within our research group. A previous study by Gammell et al., (Gammell et
al. 2007) identified components of the miR-23a~27a~24-2 cluster to be down-regulated
when subjected to hypothermic growth conditions. Stable depletion of all three miRNA
members improved both CHO cell growth and volumetric productivity (Data not shown)
in a CHO-SEAP mixed population. We sought to dissect the role that miR-23 plays in
mediating the enhanced rCHO-SEAP phenotype by diverting its actions away from its
endogenous targets. Although specified as miR-23b in this instance, the mature sequence
of miR-23b and miR-23a only differ by a single nucleotide (Appendix Figure A5), which
lies outside the miRNA seed region. miRNA sponges have been shown to sequester entire
seed families (Kluiver et al. 2012a) as miRNA recognition elements (MREs) are fully
complementary to the mature seed sequence thus effectively targeting both miR-23a and
miR-23b in our stable cell line (here in referred to as miR-23).
182
3.3.2 miR-sponge vector design and stable CHO cell line generation
Two individual miRNA sponge vectors were generated for miR-23 and miR-23b* using
the same miR-sponge construct described for the stable depletion of miR-7 (Sanchez et
al. 2013b). Both mature miRNA sequences are conserved across human (hsa) and CHO
(cgr) (Fig. 3.30 A). The miRNA sponge design for miR-23 and miR-23b* follows the
same vector construct shown in Figure 3.30 B. Molecular cloning and stable cell line
generation was carried out as described for the miR-34 sponge (section 3.2).
The miRNA sponge sequences for miR-23 and miR-23b* are shown in Table 3.3 and
included the same base-pair mis-matches across nucleotides 9-12 (miR-23) and 10-13
(miR-23b*). Molecular cloning gels for construct preparation and sequence information
can be found in the appendix (Appendix fig. A6 and A7, respectively).
Figure 3.30: miR-23b and -23b* sequence conservation and sponge design
Figure 3.30: A) Demonstrates the conservation of miR-23 and miR-23b* across the species of Homo
sapiens (hsa), Mus musculus (mmu) and Cricetulus griseus (cgr) including the dogmatic “seed” region
(yellow box). B) Schematic diagram of the miR-sponge vector backbone expressing dEGFP and
under Hygromycin selection with XhoI and EcoRI restriction sites for insertion of miR-23b*
sequence carrying fully complementary concatemeric repeats (5’-3’) to each respective microRNAs
(5’-3’) including base-pair mismatch “Wobble” at nucleotides 9-12 (miR-23) and 10-13 (miR-23b*)
(red).
EcoRIXhoI EcoRI ATG
hsa-miR-23b AUCACAUUGCCAGGGAUUACCmmu-miR-23b AUCACAUUGCCAGGGAUUACCcgr-miR-23b AUCACAUUGCCAGGGAUUACC
hsa-miR-23b* UGGGUUCCUGGCAUGCUGAUUUmmu-miR-23b* GGGUUCCUGGCAUGCUGAUUUcgr-miR-23b* GGGUUCCUGGCAUGCUGAUU
TCTAGACTCGAG.....GGTAATCCGCATCAATGTGATtcatcGGTAATCCGCATCAATGTGAT......GAATTCTCTAGA
3’-CCATTAGG GTTACACTA-5’GACC
XhoI EcoRI
TCTAGACTCGAG.....AAATCAGCCGTACAGGAACCCAtcatcAAATCAGCCGTACAGGAACCCA....GAATTCTCTAGAXhoI EcoRI
TACG3’-TTTAGTCG GTCCTTGGGT-5’
miR-23b
miR-23b*
A
B
183
Table 3.3: miRNA sponge sequence design for miR-23b/-23b*
5’-3’ microRNA Sponge sequence
miR
Sponge
XbaI
Res.
XhoI
Res.
Sponge seq. EcoRI
Res.
XbaI
Res.
miR-
23b*
TCTA
GA
CTCGA
G
acgcgAAATCAGCCGTACAGGAACCCA
tcatcAAATCAGCCGTACAGGAACCCAa
ccgtaAAATCAGCCGTACAGGAACCCAa
cgagAAATCAGCCGTACAGGAACCCAt
catc
GAA
TTC
TCTA
GA
miR-
23b
TCTA
GA
CTCCG
A
acgcgGGTAATCCGCATCAATGTGATtca
tcGGTAATCCGCATCAATGTGATaccgta
GGTAATCCGCATCAATGTGATacgagG
GTAATCCGCATCAATGTGATtcatc
GAA
TTC
TCTA
GA
3.3.3 Investigation of GFP fluorescence
Binding specificity of each miRNA sponge was assessed for its target miRNA. All three
miRNA sponges were transfected (NC-spg, 23b-spg and 23b*-spg) with their appropriate
pre-miR and separately with a negative control mimic (pre-miR-NC). A ~90% (p ≤ 0.001)
reduction in the levels of GFP expression was achieved when pre-miR-23b* and pre-miR-
23b*-mod (pre-miR-23b) were transfected into the appropriate CHO-sponge population
with the NC-sponge demonstrating little specificity (Fig. 3.31 A and B).
One point to note was the nomenclature for pre-miR-23b referred to as pre-miR-23b*-
mod. This “mod” implies a modification that we sought when ordering the commercial
hairpins. Normal pre-miR-23b and pre-miR-23b* take the form of siRNA-like constructs,
that being a fully complementary duplex with terminal modifications to allow selection
of the desired strand. In the case of pre-miR-23b*-mod, the hairpin structure mimics that
of the endogenous miR-23b/23b* duplex (Appendix Fig. A8) thus thermodynamically
favouring the selection of the mature miR-23b strand while generating the by-product
passenger strand miR-23b*. The modified version (pre-miR-23b*-mod) repressed GFP
expression with over 5-fold greater potency than the commercial siRNA-like pre-miR-
23b. This suggested that processing of the commercial pre-miR-23b duplex resulted in a
lower mature miR-23b copy number.
The non-specific sponge mixed pool population demonstrated an average higher
fluorescence when compared to both miR-23 and miR-23b* sponge mixed pools (Fig.
3.31 B). This reduced GFP fluorescence is likely a reflection of the repression caused by
184
endogenous miR-23/23b*, similar to the observation for the miR-34 sponge, in the
previous section.
The heterogeneity within all populations generated is evident from the various GFP
expression arising from individual clones within the mixed populations (Fig. 3.31 A).
Figure 3.31: miR-23b/23b*spg binding specificity through GFP suppression
Figure 3.31: miR-sponge specificity was carried out for the miR-23 and miR-23b* sponge through
the transient over-expression pre-miR-23b* and pre-miR-23b. miR-23b*-mod is a specifically
designed pre-miRNA that is composed of the mature miR-23b and miR-23b* and is processed
according to their natural thermodynamic properties with the miR-23b* strand being tagged with a
biotin group. Sponge specificity was determined based on inhibition of GFP expression through
monitoring with A) Fluorescent microscopy and B) Guava ExpressPlus programme. Non-specific
stable sponge was transfected in order to identify its lack of specificity. Statistical significance was
calculated using a standard student t-test (p ≤ 0.001***).
pre-miR-NC pre-miR-23b pre-miR-23b-mod pre-miR-23b*
0
500
1000
1500
2000
2500
GF
P E
xp
ress
ion
pre-miR treatment
miR-spgCHO-1.14 binding efficiency
GFP
miR-NC spg miR-23b spg miR-23b* spg
******
***
A
B
185
3.3.4 Analysis in mixed populations
The following section assesses the impact both stable miRNA depletion of miR-23 and
miR-23b* had in both rCHO populations (SEAP and IgG) under various bioprocess
conditions (Batch and Fed-batch).
3.3.4.1 Impact on bioprocess phenotypes in CHO-K1-SEAP: Batch and Fed-batch
miR-23 sponge mixed population - Batch
Mixed pool rCHO-SEAP cells with stable inhibition of miR-23 exhibited no growth or
viability benefit when compared to NC-spg cells (Fig. 3.32 A and B). However, an
increase ranging from ~30-50% (p ≤ 0.001) in volumetric SEAP activity was observed
(Fig. 3.32 C).
Figure 3.32: Batch process of mixed pool CHO-SEAP cells stably depleting miR-23
Figure 3.32: CHO-SEAP cells engineered to stably deplete miR-23 in a mixed pool population were
evaluated for bioprocess relevant phenotypes A) Cellular proliferation, B) Cellular viability, C)
Volumetric SEAP yield/mL and D) specific SEAP productivity/cell over the course of a 6 day batch
process and compared to a stable mixed pool population expressing a non-specific control sponge
(miR-NC spg). Statistical analysis was calculated using a standard student t-test (p ≤ 0.01** and ≤
0.001***) with an n = 3 of experimental replicates.
0
0.5
1
1.5
2
2.5
3
3.5
4
4.5
5
0 1 2 3 4 5 6 7
Cel
l N
um
ber
s/m
L (
x1
06)
Day
050
100150200250300350400450500
0 1 2 3 4 5 6 7
SE
AP
Un
its/
mL
Day
0
0.00005
0.0001
0.00015
0.0002
0.00025
0.0003
0 2 4 6
Cel
l sp
ecif
ic P
rod
uct
ivit
y
Day
miR-23b spongeMP
miR-NCspg MPmiR-23bspg MP
0
20
40
60
80
100
120
0 1 2 3 4 5 6 7
Via
bil
ity %
Day
0
0.5
1
1.5
2
2.5
3
0 1 2 3 4 5 6 7
No
rma
lise
d P
rod
uct
ivit
y (
x1
0-4
)
Day
A B
C D
**
***
***
186
miR-23 sponge mixed population – Fed-batch
miR-23 depletion conferred no growth benefit, however, addition of the feed improved
maximal cell density by 2-fold when compared to batch culture (Fig. 3.33 A). Viability
was extended a further 3 days above 80% for both miR-23/-NC sponge pools with no
additional benefit observed in the case of miR-23 depletion (Fig. 3.33 B).
Figure 3.33: Fed-batch process of mixed pool CHO-SEAP cells stably depleting miR-
23
Figure 3.33: Stable miR-23 sponge mixed pool populations were cultured in a fed-batch format over
the course of an 8 day assay. Nutrient Feed (CD CHO Efficient Feed A) was initially inoculated at
15% (v/v) on day 0 and subsequent feeds of 10% (v/v) every 72 hs (Feed signified in orange). Stable
inhibition of miR-23b was assessed for its influence on A) Cellular proliferation, B) Cell viability, C)
Volumetric SEAP activity and D) Specific SEAP activity. Statistical significance was calculated using
a standard student t-test (p ≤ 0.05* and ≤ 0.01**) with an experimental replication of n = 3.
A ~50% (p ≤ 0.05 and p ≤ 0.01 at various time points) increase in volumetric SEAP
activity was achieved across various time points for miR-23 depleted mixed pools (Fig.
3.33 C). The maximum SEAP units calculated for the mixed pool miR-23 sponge in the
0
2
4
6
8
10
12
0 2 4 6 8 10
No
rma
lise
d P
rod
uct
ivit
y (
x1
0-5
)
Day
40
50
60
70
80
90
100
110
0 2 4 6 8 10
Via
bili
ty %
Day
miR-spgCHO-K1-SEAP Fed-Batch
miR-NC spgMP
miR-23b spgMP
0.00E+00
2.00E+06
4.00E+06
6.00E+06
8.00E+06
1.00E+07
1.20E+07
0 2 4 6 8 10
Cell
Nu
mb
ers/
mL
Day
miR-23b sponge MP FB
0
50
100
150
200
250
300
350
0 2 4 6 8 10
SE
AP
Un
its/
mL
Day
miR-23b spgCHO-K1-SEAP MP Fed-Batch
miR-NC spg MP
miR-23b spg MP
*
**
***
*
**
***
40
50
60
70
80
90
100
110
0 2 4 6 8 10
Via
bil
ity
%
Day
miR-spgCHO-K1-SEAP Fed-Batch
miR-NC spgMP
miR-23b spgMP
A B
C D
Cel
l N
um
ber
s/m
L (
x1
06)
2
4
6
8
10
12
187
fed-batch process was lower when compared to batch. This reduced SEAP activity
between miR-23 depleted mixed pools under different culture conditions could
potentially be due to the 2-fold increase in cell growth resulting in an increased demand
on cellular energy metabolism to be pledged toward cellular division rather than
productivity. Finally, there was a range of 30-80% enhanced normalised SEAP
productivity observed for miR-23 depleted mixed pools compared to controls (Fig. 3.33
D).
miR-23b* sponge mixed population - Batch
Stable depletion of miR-23b* in an rCHO-SEAP mixed pool demonstrated an average
increase of 30-40% in the maximal cell density on day 4 of culture with no benefit to
culture viability when compared to NC-spg pools (Fig. 3.34 A and B). A trend towards
an enhanced growth phenotype was observed over the course of several replicates
although not significant. A 40-50% (p ≤ 0.01, p ≤ 0.001) reduced volumetric SEAP
activity was observed when compared to non-specific control sponge cells (Fig. 3.34 C).
Normalised productivity was observed to be reduced by ~60% (Fig. 3.34 D). This
reduction in volumetric/normalised productivity could potentially be due to the enhanced
proliferation phenotype diverting cellular resources from productivity.
188
Figure 3.34: Batch process of mixed pool CHO-SEAP cells stably depleting miR-
23b*
Figure 3.34: Mixed populations engineered to stably deplete miR-23b* (miR-23b* spg MP) were
cultured in a batch process over a 6 day time course with a control population expressing a non-
specific miRNA sponge. Phenotypes recorded were A) Cellular proliferation, B) Cellular viability, C)
Volumetric SEAP activity and D) Specific SEAP activity. Statistical significance was calculated based
on a standard student t-test (p ≤ 0.01** and ≤ 0.001***) and an experimental repeat of n = 3.
miR-23b* sponge mixed population – Fed-batch
When the same mixed pool populations were evaluated in a fed-batch format, a 60-70%
(p ≤ 0.01, p ≤ 0.001) increase in maximal cell density was achieved (Fig. 3.35 A) with no
impact on cellular viability.
Volumetric SEAP output was observed to be reduced by 40-75%, as was described for
batch culture (Fig. 3.35 B). However, at the later time points in culture, volumetric
productivity achieved comparative levels to that of sponge controls (Fig. 3.35 B). This
gradual increase in SEAP activity per mL is potentially attributed to the increase in the
viable cell population. A consistent reduction in normalised productivity of 50-80%
throughout various time points was observed (Fig. 3.35 C).
0
1
2
3
4
5
6
7
0 1 2 3 4 5 6 7
Cel
l N
um
ber
s/m
L (
x1
06)
Day
A
0
20
40
60
80
100
120
0 1 2 3 4 5 6 7
Via
bil
ity
%
Day
B
0
50
100
150
200
250
300
350
0 1 2 3 4 5 6 7
SE
AP
Un
its/
mL
Day
0
5
10
15
20
25
30
0 1 2 3 4 5 6 7
No
rma
lise
d P
ord
uct
iviy
(x1
0-5
)
Day
0
0.00005
0.0001
0.00015
0.0002
0.00025
0.0003
0 1 2 3 4 5 6 7C
ell
Sp
ecif
ic P
rod
uct
ivit
y
Day
miR-23b* sponge MP
miR-NC spg MP
miR-23b* spg MP
***
***
**
DC
189
Figure 3.35: Fed-Batch process of mixed pool CHO-SEAP cells stably depleting
miR-23b*
Figure 3.35: CHO-SEAP mixed populations engineered with a miRNA sponge specific for miR-23b*
(miR-23b* spg MP) were cultured over a 6 day fed-batch period alongside a non-specific control
sponge population (miR-NC spg MP). Cultured were fed initially with a nutrient rich feed on day 0
(15% v/v) and subsequently every 3 days (10% v/v). Bioprocess phenotypes assessed were A) Cellular
proliferation, B) Cell viability, C) Volumetric SEAP activity and D) Specific SEAP activity. Statistical
significance was calculated using a standard student t-test (p ≤ 0.01** and ≤ 0.001***) with an
experimental repeat of n = 3.
3.3.4.2 Impact of miR-23 and -23b* depletion on Mab producing clones
Both miRNA sponge mixed populations were next evaluated in a cell line secreting an
IgG antibody.
miR-23 sponge mixed population - Batch
The maximal cell density achieved was increased~2-fold (p ≤ 0.001) in miR-23 depleted
CHO-1.14 cells with an improved viability of 15-20% (p ≤ 0.001) when compared to
0
1
2
3
4
5
6
7
8
9
10
0 1 2 3 4 5 6 7
Cel
l N
um
ber
s/m
L (
x1
06)
Day
0
20
40
60
80
100
120
140
160
180
200
0 1 2 3 4 5 6 7
SE
AP
Un
its/
mL
Day
0
0.000005
0.00001
0.000015
0.00002
0.000025
0.00003
0.000035
0.00004
0.000045
0 1 2 3 4 5 6 7
No
rma
lise
d P
rod
uct
ivit
y (
x1
0-6
)
Day
miR-NC spg MP
miR-23b* spg MP
0
0.5
1
1.5
2
2.5
3
3.5
4
4.5
0 1 2 3 4 5 6 7
No
rma
lise
d P
rod
uct
ivit
y (
x1
0-5
)
Day
A
B C
**
*****
***
190
control sponge cells (Fig. 3.36 A and B). When volumetric IgG concentration was
determined, a ~60% (p ≤ 0.001) reduction early in culture but only ~30% (p ≤ 0.001) in
late culture was observed when compared to non-specific sponge controls (Fig. 3.36 C).
The gradual increase could be a result of the higher viable density of producing cells,
especially when productivity on a per cell basis was determined to be reduced by ~60%
across all time points (Fig. 3.36 D).
Figure 3.36: Batch process of a CHO-1.14 mixed pool population with stable
depletion of miR-23
Figure 3.36: CHO-1.14 cells engineered with a sponge specific for miR-23 and a non-specific control
sponge were cultured in a batch process over 6 days. Phenotypes assessed were A) Cellular
proliferation, B) Cellular viability with an 80% viability threshold flagged with a red broken line, C)
Volumetric IgG concentration and D) Specific IgG concentration. Statistical analysis was determined
using a standard student t-test (p ≤ 0.01** and ≤ 0.001***) with an experimental repeat of n = 3.
miR-23b* sponge mixed population - Batch
CHO-1.14 mixed pool populations demonstrated a similar phenotype to what was
observed in CHO-SEAP mixed pools. An increase in CHO cell growth of 20-50% (p ≤
0.01, p ≤ 0.001) was observed for miR-23b* sponge cells when compared to non-specific
0.00E+00
5.00E+05
1.00E+06
1.50E+06
2.00E+06
2.50E+06
3.00E+06
0 2 4 6 8
Cel
ls/m
L
Day
miR-NC spg
miR-23b
0
0.5
1
1.5
2
2.5
3
0 1 2 3 4 5 6 7
Cel
l N
um
ber
s/m
L (
x1
06)
Day
******
**
A
50
55
60
65
70
75
80
85
90
95
100
0 2 4 6
Via
bil
ity
%
Day
B
***
0
0.5
1
1.5
2
2.5
3
3.5
4
0 1 2 3 4 5 6 7
IgG
ng
/mL
(x1
05)
Day
***
***
***
0
10
20
30
40
50
60
70
80
0 1 2 3 4 5 6 7
IgG
-p
g/c
ell/
day
Day
C
D
191
pools (Fig. 3.37 A). Additionally, an increase in cellular viability of 10-15% (p ≤ 0.001)
was observed on day 6 of culture (Fig. 3.37 B). When volumetric IgG was assessed, a 30-
40% (p ≤ 0.001) decrease was determined across all time points with a higher decrease in
cell-specific productivity of 50-60% (Fig. 3.37 C and D).
Figure 3.37: Batch culture for IgG-secreting miR-23b* sponge stable CHO-1.14
Figure 3.37: Stable mixed pool CHO-1.14 population engineered with a sponge specific for miR-23b*
and a non-specific sponge sequence were cultured over a 6 day batch process. Bioprocess phenotypes
under investigation were A) Cellular proliferation, B) Cellular viability with an 80% threshold
marked with a red broken line, C) Volumetric IgG concentration and D) Specific IgG concentration.
Statistical significance was calculated using a standard student t-test (p ≤ 0.01** and ≤ 0.001***) and
an experimental repeat of n = 3.
3.3.5 Analysis of single cell clones in 24-well plate
Clones from all mixed populations (miR-23, miR-23b* and miR-NC) were selected based
on a medium-high level of GFP fluorescence to ensure that clones selected were
expressing the miR-sponge constructs to a degree sufficient enough to meet the minimum
threshold levels to achieve saturated target miRNA diversion (Fig. 3.38).
0
10
20
30
40
50
60
70
80
0 1 2 3 4 5 6 7
IgG
-p
g/c
ell/
day
Day
0
0.5
1
1.5
2
2.5
0 1 2 3 4 5 6 7
Cel
ls/m
L (
x1
06)
Day
A
50
55
60
65
70
75
80
85
90
95
100
0 2 4 6
Via
bil
ity
%
Day
Viability %
miR-NC spg
miR-23b* spg50
55
60
65
70
75
80
85
90
95
100
0 2 4 6
Via
bil
ity
%
Day
B
***
***
***
***
0
0.5
1
1.5
2
2.5
3
3.5
4
0 1 2 3 4 5 6 7
IgG
ng
/mL
(x1
05)
Day
C
***
***
***
D
192
Figure 3.38: GFP distribution of miR-sponge clones isolated using FACS
Figure 3.38: miR-sponge clones were sorted based on a medium-high range of GFP intensity. Mixed
populations were gated based on a non-fluorescent parental CHO-K1-SEAP cell line. The average
GFP intensity is shown in the top right hand corner of the graphs. The percentage of medium-high
GFP cells is represented as a percentage compared to all cells.
58.9%
6,898
5,999
4%
3,334
3.6%
Par
enta
l C
HO
-SE
AP
miR
-NC
sp
on
ge
miR
-23
b s
po
ng
em
iR-2
3b
* s
po
ng
e
193
Before phenotypic evaluation was performed, all clones were passaged at least 3 time in
the presence of PVA and pulsed with G418 (Neomycin Phosphotransfarase) and HYG
(Hygromycin phosphotransfarase).
3.3.5.1 GFP fluorescence across clonal panels
From the panel of 24 clones in each group (miR-NC-spg, miR-23-spg and miR-23b*-spg)
in both cell lines (CHO-SEAP and CHO-1.14), GFP fluorescence was determined using
the Guava ExpressPlus programme on the GuavaTM benchtop cytometer. It revealed that
there was a spread in GFP expression intensity throughout all clonal panels ranging from
very high to low expression levels (Fig. 3.39). Similar to what was observed for the miR-
34 sponge, the average GFP intensity of both miR-23 and miR-23b* panels was reduced
compared to controls and similarly observed in FACS readouts of miR-23/23b* sponge
mixed pools (Fig. 3.38). It was also observed that the reduced GFP expression in the case
of the miR-23b* panel was of a lower magnitude when compared to the control panel as
opposed to miR-23 (Fig. 3.39). This could potentially be attributed to the nature of miR*
strands being predominantly less abundant in addition to suggesting a potential lesser
phenotype. The range in GFP intensity across the panel of clones is potentially attributed
to plasmid integration site within the genome in addition to copy number and not as a
result of transfection efficiency. Efficiency of transfection would have been assessed 24
h post-transfection (Data not shown) with a ~70% efficiency in the case of both miR-NC
and miR-34 sponge mixed pools.
194
Figure 3.39: GFP expression across miR-spg clonal panel (miR-NC, miR-23b and miR-23b*) in both CHO-K1-SEAP and CHO-
1.14 cells
Figure 3.39: GFP expression for each individual clone from all stable miR-sponge populations (miR-NC spg, miR-23b spg and miR-23b* spg) in two different
CHO cell lines (CHO-K1-SEAP and CHO-1.14). The Median GFP fluorescence for each panel is represented as a broken red line. GFP intensity was analysed
using the Guava ExpressPlus programme on a Guava EasyCyte benchtop cytometer.
A
B
0
500
1000
1500
2000
2500
3000
GF
P E
xp
ress
ion
miR-NC sponge clonesCHO-1.14
GFP
0
500
1000
1500
2000
2500
3000
GF
P E
xp
ress
ion
miR-23b sponge clonesCHO-1.14
GFP
0
500
1000
1500
2000
2500
3000
GF
P E
xp
ress
ion
miR-23b* sponge clonesCHO-1.14
GFP
miR-NC spgCHO-1.14 miR-23b spgCHO-1.14 miR-23b* spgCHO-1.14
Median GFP = 788.525 Median GFP = 528.55 Median GFP = 738.86
0
200
400
600
800
1000
1200
1400
GF
P E
xp
ress
ion
miR-NC sponge clonesCHO-K1-SEAP
GFP
0
200
400
600
800
1000
1200
1400
GF
P E
xp
ress
ion
miR-23b sponge clonesCHO-K1-SEAP
GFP
0
200
400
600
800
1000
1200
1400
GF
P E
xp
ress
ion
miR-23b* sponge clonesCHO-K1-SEAP
GFP
Median GFP = 242.325 Median GFP = 230.115Median GFP = 437.985
miR-NC spgCHO-K1-SEAP miR-23b spgCHO-K1-SEAP miR-23b* spgCHO-K1-SEAP
195
3.3.5.2 Impact on bioprocess phenotypes across CHO-K1-SEAP clonal panels
In the mixed pool populations under various culture conditions, stable depletion of miR-
23 in our rCHO-SEAP cell line demonstrated no impact on CHO cell growth or viability
yet significantly improved volumetric SEAP production due to enhanced specific
productivity. On the other hand, miR-23b* depletion appeared to have a subtle beneficial
effect on CHO cell growth at the expense of specific productivity. Clones isolated from
both panels were evaluated on a clonal level as a means of assessing the gross phenotype
as a function of the average panel performance.
miR-23 sponge clonal panel - Batch
After adaption to suspension culture, clones were grown in a 1 mL volume of serum-free
medium and cultured in a 24-well suspension plate. The original panel of 24 clones,
previously assessed for GFP fluorescence, was reduced to a panel of 14 clones due to loss
during the adaption process as well as a contamination event over the course of several
passages. The same occurred for miR-NC sponge clones of which 15 clones were
recovered.
Both clonal panels were cultured over the course of a 6 day batch period and sampled on
days 2, 4 and 6.
The average growth was not observed to differ when clones were assessed on day 2 (Fig.
3.40Day 2), as observed in the mixed pools. However, clones stably expressing the miR-23
sponge decoy achieved half the maximal cell density on day 4 when compared to the
miR-NC sponge panel (Fig. 3.40Day 4), despite a subset of clones reaching “normal” cell
densities. Cell density was further reduced by ~70% on day 6, potentially attributed to the
25% drop in cell viability at this late culture time point (Fig. 3.40Day 6, Fig. 3.41).
196
Figure 3.40: CHO-SEAP miR-23 sponge clonal panel – Cell numbers/mL
Figure 3.40: CHO-SEAP clonal panel demonstrating stable depletion of miR-23 and assessed for
their growth phenotype in 1 mL suspension culture in a 24-well plate format and sampled at days 2,
4 and 6 in a batch process. Average cell numbers calculated across the entire clonal panels is
represented by the broken red line with the miR-NC-spg panel represented in dark blue and miR-
23b spg panel in light blue. Yellow boxes highlight the individual sponge clones that were
subsequently selected from each group for scale up and an n = 2.
0.00E+00
2.00E+05
4.00E+05
6.00E+05
8.00E+05
1.00E+06
1.20E+06
1.40E+06
1 17 15 10 6 22 9 24 2 13 11 21 8 19 20
Via
ble
Cell
s/m
L
Clone
miR-NC spg Cells/mL Day 2
Cells/mL
0.00E+00
2.00E+05
4.00E+05
6.00E+05
8.00E+05
1.00E+06
1.20E+06
1.40E+06
21 7 4 12 6 2 14 22 10 3 23 11 24 9
Via
ble
Cell
s/m
L
Clone
miR-23b spg Cells/mL Day 2
Cells/mL
Ave Cells/mL = 7.63 x 105 Ave Cells/mL = 7.67 x 105
0.00E+00
1.00E+06
2.00E+06
3.00E+06
4.00E+06
5.00E+06
6.00E+06
7.00E+06
6 15 17 24 10 22 2 9 19 13 8 1 21 11 20
Via
ble
Cell
s/m
L
Clone
miR-NC spg Cells/mL Day 4
Cells/mL
0.00E+00
1.00E+06
2.00E+06
3.00E+06
4.00E+06
5.00E+06
6.00E+06
7.00E+06
4 6 7 14 12 11 21 3 2 22 24 10 23 9
Via
ble
Cell
s/m
L
Clone
miR-23b spg Cells/mL Day 4
Cells/mL
Ave Cells/mL = 4.98 x 106 Ave Cells/mL = 2.23 x 106
Day 2
Day 4
0.00E+00
1.00E+06
2.00E+06
3.00E+06
4.00E+06
5.00E+06
6.00E+06
9 24 17 6 10 8 21 13 11 19 22 15 1 20 2
Via
ble
Cell
s/m
L
Clone
miR-NC spg Cells/mL Day 6
Cells/mL
Day 6
0.00E+00
1.00E+06
2.00E+06
3.00E+06
4.00E+06
5.00E+06
6.00E+06
23 24 3 11 14 22 12 2 10 9 21 4 7 6
Via
ble
Cell
s/m
L
Clone
miR-23b spg Cells/mL Day 6
Cells/mL
Ave Cells/mL = 1.01 x 106Ave Cells/mL = 3.55 x 106
Clone
miR-23 spg panelmiR-NC spg panel
Cel
l N
um
ber
s/m
L (
x1
05)
Cel
l N
um
ber
s/m
L (
x1
06)
Cel
l N
um
ber
s/m
L (
x1
06)
Day 2
Day 4
Day 6
1
0
2
3
4
5
6
0
0
1
2
2
4
6
8
10
12
14
3
4
5
6
7
197
Figure 3.41: Day 6 cell viability of miR-23b CHO-SEAP sponge panel
Figure 3.41: As cellular viability was comparable between both clonal panels on day 2 and 4, for both
the miR-23 spg and miR-NC spg panel only the viability of each individual clone form both panel for
day 6 of culture is presented above. The average viability across each panel is represented as a broken
red line with the corresponding value enclosed. Clones ultimately selected for scale up are highlighted
in orange boxes and an n = 2.
An increase of 3.1, 2.9 and 2.5-fold (p ≤ 0.001) in SEAP production per mL was observed
on days 2, 4 and 6, respectively in the miR-23 depleted panel despite the negative impact
on cell density (Fig. 3.42). Although some clones from the NC-sponge group possessed
a high level of SEAP production, 11 of the 15 miR-23 sponge clones outperformed the
highest performing control clone. In the miR-23 sponge panel, average volumetric SEAP
activity was observed to be 2.5-fold increased on day 4. This translated to a 5.8-fold (p ≤
0.001) increase in normalised productivity due to the reduced cell numbers (Fig. 3.40Day
4, 3.42Day 4 and 3.43Day 4). This suggests that although miR-23 may be contributing to
enhanced specific productivity, the extent to which an individual cell will respond relies
on various cellular processes such as proliferation, an energetically demanding process.
0
10
20
30
40
50
60
70
80
90
11 9 8 24 17 10 20 6 22 19 1 13 21 15 2
Via
bil
ity
%
Clone
miR-NC spg Viability (Day 6)
0
10
20
30
40
50
60
70
80
90
9 24 23 3 11 14 22 12 2 10 21 7 4 6
Via
bil
ity
%
Clone
miR-23b spg Viability (Day 6)
Ave viability =56% Ave viability =26%
miR-NC spg panel miR-23 spg panelClone
Day 6
198
Figure 3.42: miR-23 sponge clonal panel – Volumetric SEAP activity
Figure 3.42: Individual volumetric SEAP activity is presented above for both miR-23 sponge and
NC-sponge clones with the average performance across both panels represented as a red broken line.
Orange boxes highlight the individual sponge clones that were subsequently selected from each group
for scale up. Statistical significance was determined using a standard student t-test (p ≤ 0.001***), n
= 2.
0
200
400
600
800
1000
6 4 23 2 21 12 10 22 14 7 9 11 3 24
SE
AP
Un
its/
mL
Clone
miR-23b spg SEAP Day 6
SEAP/mL
0
100
200
300
400
500
600
700
800
900
4 6 23 2 21 12 10 7 14 22 11 3 9 24
SE
AP
Un
its/
mL
Clone
miR-23b spg SEAP Day 4
SEAP/mL
0
50
100
150
200
250
300
4 6 12 7 2 21 10 23 14 22 9 11 3 24
SE
AP
Un
its/
mL
Clone
miR-23b spg SEAP Day 2
SEAP/mL
-50
0
50
100
150
200
250
300
11 20 22 21 15 8 2 6 1 19 12 10 9 17 24
SE
AP
Un
its/
mL
Clone
miR-NC SEAP Day2
SEAP/mL
Ave SEAP = 36.15 Ave SEAP = 110.651
0
100
200
300
400
500
600
700
800
900
11 2 21 22 1 8 20 15 6 19 9 13 10 17 24
SE
AP
Un
its/
mL
Clone
miR-NC spg SEAP Day 4
SEAP/mL
Ave SEAP = 150.754 Ave SEAP = 437.31
0
200
400
600
800
1000
11 8 2 20 21 15 1 22 6 13 19 9 10 24 17
SE
AP
Un
its/
mL
Clone
miR-NC spg SEAP Day 6
SEAP/mL
Ave SEAP = 221.173 Ave SEAP = 542.841
Clone
miR-23 spg panelmiR-NC spg panel
Day 6
Day 4
Day 2
***
***
***
199
Figure 3.43: miR-23 sponge clonal panel – specific SEAP productivity
Figure 3.43: Specific SEAP productivity for all clones in both miR-23 and miR-NC sponge panels.
Average specific productivity is represented as a red broken line. Orange boxes highlight the
individual sponge clones that were subsequently selected for scale up. Statistical analysis was
calculated using a standard student t-test (p ≤ 0.05*, ≤ 0.01** and ≤ 0.001***), n = 2. The outlier,
miR-23 clone 6 on day 6, is at a value of 4.98 x 10-2.
Finally, clones selected from both miR-23 sponge and the miR-NC sponge panels were
chosen based on volumetric SEAP production, highlighted in orange boxes (Fig. 3.42).
From the miR-23 sponge panel, the four best producing clones were chosen for scale up
(miR-23 sponge 2, 4, 6 and 23). A similar approach was taken for the selection of the
miR-NC sponge clones (miR-NC sponge 2 and 11) except that miR-NC sponge clone 20
0.00E+00
5.00E-04
1.00E-03
1.50E-03
2.00E-03
2 20 1 11 15 8 21 22 6 19 13 10 9 24 17
SE
AP
Un
its/
cell
0.00E+00
5.00E-03
1.00E-02
1.50E-02
2.00E-02
6 4 7 21 9 10 2 12 22 14 23 11 3 24
SE
AP
Un
its/
cell
0.00E+00
5.00E-05
1.00E-04
1.50E-04
2.00E-04
2.50E-04
3.00E-04
3.50E-04
20 11 21 8 22 2 15 19 6 13 1 9 10 17 24
SE
AP
Un
its/
cell
Clone
miR-NC Qp Day 2
SEAP/cell
0.00E+00
5.00E-05
1.00E-04
1.50E-04
2.00E-04
2.50E-04
3.00E-04
3.50E-04
9 10 23 4 6 12 2 7 21 22 14 11 3 24
SE
AP
Un
its/
cell
Clone
miR-23b spg panel Day 2
SEAP/cell
0.00E+00
1.00E-04
2.00E-04
3.00E-04
4.00E-04
5.00E-04
6.00E-04
7.00E-04
8.00E-04
9.00E-04
23 10 9 2 22 6 4 21 12 14 7 11 3 24
SE
AP
Un
its/
cell
Clone
miR-23b spg panel Day 4
SEAP/cell
0.00E+00
1.00E-04
2.00E-04
3.00E-04
4.00E-04
5.00E-04
6.00E-04
7.00E-04
8.00E-04
9.00E-04
23 10 9 2 22 6 4 21 12 14 7 11 3 24
SE
AP
Un
its/
cell
Clone
miR-23b spg panel Day 4
SEAP/cell
-1.00E-04
1.00E-18
1.00E-04
2.00E-04
3.00E-04
4.00E-04
5.00E-04
6.00E-04
7.00E-04
8.00E-04
9.00E-04
20 11 21 2 1 8 22 19 15 13 6 9 10 17 24
SE
AP
Un
its/
cell
Clone
miR-NC Qp Day 4
SEAP/cell
Ave =6.64 x 10-5 SEAP Units/cell Ave =1.62 x 10-4 SEAP Units/cell
Ave =4.41 x 10-5 SEAP Units/cell Ave =2.38 x 10-4 SEAP Units/cell
Ave =6.84 x 10-3 SEAP Units/cellAve =1.02 x 10-4 SEAP Units/cell
Day 4
Day 6
miR-23 spg panelmiR-NC spg panelClone
**
***
*
2.4-fold
5.8-fold
66.7-fold
SE
AP
Un
its/
Cel
l (x
10
-4)
SE
AP
Un
its/
Cel
l (x
10
-4)
SE
AP
Un
its/
Cel
l (x
10
-2)
Day 2
Day 4
Day 6
0
0.5
1
1.5
2
2.5
3.5
3
0
0
1
2
3
4
5
6
7
8
9
0.5
1
1.5
2
200
was also selected based on its high normalised productivity as well as its high volumetric
productivity.
miR-23b* sponge clonal panel - Batch
In 1 mL suspension culture, the miR-23b* sponge clonal panel did not exhibit a beneficial
growth phenotype as was suggested from previous characterisation in mixed pools.
Although some individual clones did demonstrate an enhanced growth phenotype early
into batch culture (miR-23b* spg clone 18 and 22, Fig. 3.44Day 2), on average across the
entire clonal panel, there was no growth benefit evident with clones behaving similarly
to that of the miR-NC sponge panel. Furthermore, similar to what was observed for the
miR-23 sponge panel, there was a 50% reduction in the maximal cell density achieved by
miR-23b* sponge clones (Fig. 3.44Day 4+6). However, unlike the observation from the
miR-23 sponge panel, cell viability was similar when compared to the control panel (Fig.
3.45).
201
Figure 3.44: CHO-SEAP miR-23b* sponge clonal panel – Cell numbers/mL
Figure 3.44: The individual performance of each clone derived from both miR-23b* sponge and miR-
NC sponge mixed pools when cell growth was evaluated over a batch culture process. Average
performance across all clones is represented by the broken red line with each clone that was
ultimately selected for further validation highlighted with an orange box.
0.00E+00
2.00E+05
4.00E+05
6.00E+05
8.00E+05
1.00E+06
1.20E+06
1.40E+06
1.60E+06
1 17 15 10 6 22 9 24 2 13 11 21 8 19 20
Via
ble
Cell
s/m
L
Clone
miR-NC spg Cells/mL Day 2
Cells/mL
0.00E+00
2.00E+05
4.00E+05
6.00E+05
8.00E+05
1.00E+06
1.20E+06
1.40E+06
1.60E+06
22 18 23 6 24 2 11 8 5 13 10 12 4 3 20
Cell
Nu
mb
ers/
mL
Clone
miR-23b* spgCHO-K1-SEAP Cells/mL
Day 2
0.00E+00
1.00E+06
2.00E+06
3.00E+06
4.00E+06
5.00E+06
6.00E+06
7.00E+06
6 15 17 24 10 22 2 9 19 13 8 1 21 11 20
Via
ble
Cell
s/m
L
Clone
miR-NC spg Cells/mL Day 4
Cells/mL
0.00E+00
1.00E+06
2.00E+06
3.00E+06
4.00E+06
5.00E+06
6.00E+06
7.00E+06
22 18 23 11 6 8 24 12 5 10 13 2 4 3 20
Cell
Nu
mb
ers/
mL
Clone
miR-23b* sponge panel Day 4
Day 4
0.00E+00
1.00E+06
2.00E+06
3.00E+06
4.00E+06
5.00E+06
6.00E+06
9 24 17 6 10 8 21 13 11 19 22 15 1 20 2
Via
ble
Cell
s/m
L
Clone
miR-NC spg Cells/mL Day 6
Cells/mL
0.00E+00
1.00E+06
2.00E+06
3.00E+06
4.00E+06
5.00E+06
6.00E+06
18 24 23 10 11 6 13 8 12 5 20 3 22 4 2
Cell
Nu
mb
ers/
mL
Clone
miR-23b* sponge panel Day 6
Day 6
Ave =7.63 x 105 cells/mL Ave =7.05 x 105 cells/mL
Ave =2.24 x 106 cells/mLAve =4.98 x 106 cells/mL
Ave =1.99 x 106 cells/mLAve =3.55 x 106 cells/mL
Day 6
Day 4
Day 2
Clone
miR-NC spg panel miR-23b* spg panel
Cel
l N
um
ber
s/m
L (
x1
05)
Cel
l N
um
ber
s/m
L (
x1
06)
Cel
l N
um
ber
s/m
L (
x1
06)
Day 4
Day 2
Day 6
0
0
0
2
4
6
8
10
12
14
16
1
2
3
4
5
6
7
1
2
3
4
5
6
2
2
211
11
11 20
20
20 18
18
18 8
8
8 3
3
3
202
Figure 3.45: Day 6 cell viability of miR-23b* sponge panel
Figure 3.45: Day 6 viability for the miR-23b* and miR-NC sponge panels with average viability
represented as a broken red line. No impact was observed at the earlier time points of day 2 and 4
with clones selected highlighted with an orange box.
The drastic impact previously observed on both volumetric and specific SEAP activity
for the miR-23b* sponge mixed pools was not reflected in the clonal panel. SEAP activity
remained unchanged on day 2 of culture while a ~20% reduction was observed on day 4
of culture (Fig. 3.46Day 2+4). Normalised productivity also remained unchanged on day 2
of culture while being enhanced by ~2-fold (p ≤ 0.05) on day 4 (Fig. 3.47Day 2+4).
Furthermore, although a 2-fold increase in normalised productivity was observed, this
coincided with a reduction in volumetric yield by ~20% on day 4. However, on day 6 of
culture, SEAP yield was measured to be ~35% increased with a further improved
normalised productivity of ~3.5-fold (Fig. 46Day 6, 47Day 6).
0
10
20
30
40
50
60
70
80
90
100
8 18 13 10 6 2 11 24 23 3 20 5 4 12 22
Via
bil
ity
%
Clone
miR-23b* sponge panel viability %
0
10
20
30
40
50
60
70
80
90
100
11 9 8 24 17 10 20 6 22 19 1 13 21 15 2
Via
bil
ity
%
Clone
miR-NC spg Viability (Day 6)
Ave viability =56% Ave viability =54%
miR-NC spg panel miR-23b* spg panelClone
Day 6
203
Figure 3.46: CHO-SEAP miR-23b* sponge clonal panel – Volumetric SEAP
Figure 3.46: Volumetric SEAP activity for each individual clone from both miR-23b* and miR-NC
sponge panels is represented above with the average performance across the clonal panels shown as
a broken red line. Clones selected for scale-up are highlighted in orange boxes with no statistical
significance observed for phenotypic differences calculated using a standard student t-test.
0
10
20
30
40
50
60
70
80
90
100
11 20 22 21 15 8 2 6 1 19 12 10 9
SE
AP
Un
its/
mL
Clone
miR-NC spongeCHO-K1-SEAP panel
0
10
20
30
40
50
60
70
80
90
100
18 2 3 8 22 24 10 5 4 23 12 13 11 6 20
SE
AP
Un
its/
mL
Clone
miR-23b* spongeCHO-K1-SEAP panel
0
50
100
150
200
250
300
350
400
450
500
11 2 21 22 1 8 20 15 6 19 9 13 10 17 24
SE
AP
Un
its/
mL
Clone
miR-NC spongeCHO-K1-SEAP panel (Day 4)
0
50
100
150
200
250
300
350
400
450
500
18 2 3 8 22 9 12 24 11 23 4 5 6 13 20S
EA
P U
nit
s/m
L
Clone
miR-23b* sponge SEAP/mL Day 4
0
100
200
300
400
500
600
8 2 18 3 10 24 22 12 5 4 23 6 11 13 20
SE
AP
Un
its/
mL
Clone
miR-23b* sponge SEAP/mL Day 6
0
100
200
300
400
500
600
11 8 2 20 21 15 1 22 6 13 19 9 10 24 17
SE
AP
Un
its/
mL
Clone
miR-NC sponge SEAP/mL Day 6
Ave SEAP =34.15 Ave SEAP =39
Ave SEAP =135Ave SEAP =167
Ave SEAP = 221.1 Ave SEAP = 296.7
Day 2
Day 4
Day 6
Clone
miR-NC spg panel miR-23b* spg panel
0.80-fold
1.34-fold
204
Figure 3.47: CHO-SEAP miR-23b* sponge clonal panel – Specific productivity
Figure 3.47: Specific SEAP productivity was calculated and graphed for each individual clone form
both miR-23b* and miR-NC sponge clones. Average panel performance is represented as a broken
red line again with selected clones highlighted in an orange box. Statistical significance was calculated
using a standard student t-test (p ≤ 0.05*). miR-23b* sponge clone 2 on day 6 has a value of 2.55 x 10-
3 SEAP units/cell.
Summary of miR-23 and miR-23b* clonal panels
As in indicated in mixed pools, miR-23 depleted clones demonstrated an enhanced
volumetric and normalised productivity of ~ 3-fold. Maximum cell densities achieved
were observed to be reduced by ~50% when compared to controls resulting in a further
boost in normalised productivity. miR-23b* depleted clones demonstrated no growth
benefit with the exception of a couple of clones early in culture (Day 2). Similar to the
miR-23 sponge clones, these clones also reached lower maximum cell densities which
0.00E+00
1.00E-04
2.00E-04
3.00E-04
4.00E-04
5.00E-04
6.00E-04
2 20 1 11 15 8 21 22 6 19 13 10 9 24 17
SE
AP
Un
its/
cell
0.00E+00
1.00E-04
2.00E-04
3.00E-04
4.00E-04
5.00E-04
6.00E-04
2 3 4 22 8 5 20 10 12 18 24 6 23 13 11
SE
AP
Un
its/
Cel
l
0.00E+00
5.00E-05
1.00E-04
1.50E-04
2.00E-04
2.50E-04
3.00E-04
3.50E-04
20 11 21 8 22 2 15 19 6 13 1 9 10 17 24
SE
AP
Un
its/
cell
Clone
miR-NC Qp Day 2
0.00E+00
5.00E-05
1.00E-04
1.50E-04
2.00E-04
2.50E-04
3.00E-04
3.50E-04
3 2 8 18 20 10 4 12 24 5 22 13 23 11 6
SE
AP
Un
its/
Cell
Clone
miR-23b* spgCHO-K1-SEAP panel Qp
0.00E+00
5.00E-05
1.00E-04
1.50E-04
2.00E-04
2.50E-04
3.00E-04
3.50E-04
20 11 21 2 1 8 22 19 15 13 6 9 10 17 24
SE
AP
Un
its/
cell
Clone
miR-NC Qp Day 4
0.00E+00
5.00E-05
1.00E-04
1.50E-04
2.00E-04
2.50E-04
3.00E-04
3.50E-04
3 2 8 18 4 9 12 20 24 5 11 22 6 23 13
SE
AP
Un
its/
Cell
Clone
miR-23b* spgCHO-K1-SEAP panel Qp
Ave = 6.6 x 10-5 SEAP Units/Cell Ave = 6.3 x 10-5 SEAP Units/Cell
Ave = 9.3 x 10-5 SEAP Units/CellAve = 4.4 x 10-5 SEAP Units/Cell
Ave = 1.0 x 10-5 SEAP Units/Cell Ave = 3.5 x 10-5 SEAP Units/Cell
miR-NC spg panel miR-23b* spg panelClone
2.12-fold
3.52-fold
*
2
2
2 20
20
20 11
11
11 3
3
3 8
8
8 18
18
18
205
could be contributing to the enhanced normalised productivity observed. SEAP yield was
observed to be reduced except late in culture where it was increased by 34%. Viability
was not observed to differ from control clones as opposed to miR-23 sponge clones whose
viability was reduced despite exhibiting reduced cell numbers. Clones were selected
based on volumetric productivity.
3.3.5.3 Impact on bioprocess phenotypes across CHO-1.14 clonal panels
Subsequent to the characterisation of the influence of both miR-23 sponge and miR-23b*
sponge constructs in an stable mixed pool IgG-secreting rCHO-1.14 cell line, individual
clones were sorted using FACS under the same selection criteria as was previously
applied to the selection of both miR-spg SEAP clonal panels.
miR-23 sponge clonal panel - Batch
CHO-1.14 cell clones stably expressing a sponge specific for miR-23 were cultured in a
1 mL volume of CHO-S-SFM II media in suspension batch culture over the course of 6
days. As previously described, a certain number of clones were lost over the course of
suspension adaptation and suitability to suspension culture. The improved cellular
growth, as first observed in the predecessor mixed pools, was not as potent in the case of
clones although, on average, growth was enhanced ~10% on days 2 and 4 with a more
significant ~40% (p ≤ 0.001) observed on day 6 (Fig. 3.48).
13 out of 22 miR-23 sponge clones demonstrated higher cell numbers when compared to
the average cell density of the controls. This less potent growth benefit could potentially
be attributed to the nature of the 1 mL suspension culture process of which both miR-
23/23b* sponge CHO-SEAP clones demonstrated reduced growth.
206
Figure 3.48: CHO-1.14 (IgG) miR-23 sponge clonal panel – Cell numbers/mL
Figure 3.48: Cellular proliferation for each individual clone within the miR-23 sponge panel is
represented above with the comparative miR-NC sponge panel in a CHO-1.14 cell line. The average
cell numbers/mL was calculated across each panel and is represented as a broken red line. Statistical
significance was calculated using a standard student t-test (p ≤ 0.001***). Each clones ultimately
selected for scale up is highlighted in an orange box.
0.00E+00
2.00E+05
4.00E+05
6.00E+05
8.00E+05
1.00E+06
1.20E+06
1.40E+06
20 16 9 3 21 4 13 1 6 17 14 2 11 15 8 22
Cell
Nu
mb
ers/
mL
Clone
miR-NC sponge panel CHO-1.14
0.00E+00
2.00E+05
4.00E+05
6.00E+05
8.00E+05
1.00E+06
1.20E+06
1.40E+06
11 13 7 24 14 20 12 23 3 17 18 16 10 4 5 6 15 22 9 8 21 2
Cell
Nu
mb
ers/
mL
Clone
miR-23b sponge panel CHO-1.14
0.00E+00
5.00E+05
1.00E+06
1.50E+06
2.00E+06
2.50E+06
3.00E+06
3.50E+06
4.00E+06
24 20 7 8 11 13 23 14 18 5 3 16 17 9 4 15 10 6 2 21 12 22
Cell
Nu
mb
ers/
mL
Clone
miR-23b sponge panel CHO-1.14 Day 4
0.00E+00
5.00E+05
1.00E+06
1.50E+06
2.00E+06
2.50E+06
3.00E+06
3.50E+06
4.00E+06
20 21 16 9 3 15 13 6 14 17 4 1 11 2 8 22
Cell
Nu
mb
ers/
mL
Clone
miR-NC sponge panel CHO-1.14 Day 4
0.00E+00
1.00E+06
2.00E+06
3.00E+06
4.00E+06
5.00E+06
5 24 7 11 8 9 3 4 20 6 18 10 16 13 23 14 2 17 21 15 22 12
Cell
Nu
mb
ers/
mL
Clone
miR-23b sponge panel CHO-1.14 Day 4
0.00E+00
5.00E+05
1.00E+06
1.50E+06
2.00E+06
2.50E+06
3.00E+06
3.50E+06
4.00E+06
4.50E+06
5.00E+06
16 9 15 3 6 14 4 21 13 2 8 20 1 11 17 22
Cell
Nu
mb
ers/
mL
Clone
miR-NC sponge panel CHO-1.14 Day 6
miR-23 spg panelmiR-NC spg panel
Ave = 6.08 x 105 Cells/mL
Ave = 1.97 x 106 Cells/mL
Ave = 2.05 x 106 Cells/mL Ave = 2.86 x 106 Cells/mL
Ave = 2.2 x 106 Cells/mL
Ave = 6.6 x 105 Cells/mL
***
Clone
Day 2
Day 4
Day 6
1.09-fold
1.12-fold
1.39-fold
Cell
Nu
mb
ers/
mL
(x
10
5)
Day 2
Cell
Nu
mb
ers/
mL
(x
10
6)
Cell
Nu
mb
ers/
mL
(x
10
6)
Day 4
Day 6
0
0
0
2
4
6
8
10
12
14
0.5
1
1.5
2
2.5
3
3.5
4
0.51
1.52
2.5
33.5
4
4.55
20
20
202
2
2 15
15
15 9
9
9 6
6
6 23
23
2314
14
14 17
17
17
207
The enhanced cell density at day 6 was potentially attributed to the increased cell viability
of 87.7% (p ≤ 0.001) for miR-23 sponge clones when compared to 63.4% for controls
(Fig. 3.49). The direct impact of miR-23 depletion on viability is potentially reflected in
the large percentage (18/22 or 81.1%) of clones that maintained a viability of over 80%
when compared to controls (8/16 or 50%).
Figure 3.49: CHO-1.14 (IgG) miR-23 sponge clonal panel – Cell viability %
Figure 3.49: As viability remained relatively unchanged when comparing miR-23 sponge and miR-
NC sponge clones on day 2 and 4, only day 6 cell viability was reported. The average cellular viability
across both sponge panels is represented as a broken red line. Statistical significance was calculated
using a standard student t-test (p ≤ 0.001***) with each clone ultimately selected for scale up being
highlighted in an orange box.
0
10
20
30
40
50
60
70
80
90
100
3 13 2 9 15 16 6 1 11 4 8 14 17 21 20 22
Via
bil
ity
%
Clone
miR-NC sponge CHO-1.14 Day 6 Via
0
10
20
30
40
50
60
70
80
90
100
18 5 16 3 4 8 7 9 6 10 13 23 24 20 11 2 14 17 15 21 22 12
Via
bil
ity
%
Clone
miR-23b sponge panel CHO-1.14 viability Day 6
miR-23 spg panelmiR-NC spg panel
Ave = 63.4 % Ave = 87.7 %
Clone
***
208
Figure 3.50: CHO-1.14 (IgG) miR-23 sponge clonal panel – Volumetric productivity
Figure 3.50: Volumetric IgG production is represented above for all clones across both sponge panels
with average performance shown as a red broken line over a 6 day batch culture process. Clones
selected for scale up are highlighted in orange boxes. Statistical significance was calculated using a
standard student t-test (p ≤ 0.05*).
It was observed that a significant number of individual clones completely lost IgG
production, with a higher proportion in miR-23 sponge clones. Statistical analysis was
performed on all clones. A reduction in volumetric IgG was observed on days 2 (~35%)
and 4 (~45%), and 6 (~40%, p ≤ 0.05) in miR-23 depleted clone panels (Fig. 3.50). Day
6 demonstrated the largest reduction in specific productivity (~60%, Fig. 3.51).
0.00E+00
1.00E+04
2.00E+04
3.00E+04
4.00E+04
5.00E+04
6.00E+04
7.00E+04
8.00E+04
9.00E+04
2 14 8 4 20 3 6 11 15 21 13 1 17 16 22 9
IgG
ng
/mL
Clone
miR-NC sponge CHO-1.14 IgG/mL Day 2
0.00E+00
1.00E+04
2.00E+04
3.00E+04
4.00E+04
5.00E+04
6.00E+04
7.00E+04
8.00E+04
9.00E+04
14 11 12 2 13 10 6 17 16 5 4 7 20 18 8 22 24 21 23 15 3 9
IgG
ng
/mL
Clone
miR-23b sponge CHO-1.14 IgG/mL Day 2
0.00E+00
5.00E+04
1.00E+05
1.50E+05
2.00E+05
2.50E+05
3.00E+05
3.50E+05
20 2 14 21 15 8 4 3 6 13 11 22 17 16 9 1
IgG
ng
/mL
Clone
miR-NC sponge CHO-1.14 IgG/mL Day 60.00E+00
5.00E+04
1.00E+05
1.50E+05
2.00E+05
2.50E+05
14 20 2 8 3 4 15 6 21 11 13 17 9 1 16 22
IgG
ng
/mL
Clone
miR-NC sponge CHO-1.14 IgG/mL Day 4
0.00E+00
5.00E+04
1.00E+05
1.50E+05
2.00E+05
2.50E+05
14 12 11 2 13 6 10 16 17 5 4 7 20 8 18 3 21 23 22 15 24 9
IgG
ng
/mL
Clone
miR-23b sponge CHO-1.14 IgG/mL Day 4
0.00E+00
5.00E+04
1.00E+05
1.50E+05
2.00E+05
2.50E+05
3.00E+05
3.50E+05
23 14 17 21 13 4 11 9 2 6 22 15 10 12 20 18 16 5 24 3 7 8
IgG
ng
/mL
Clone
miR-23b sponge CHO-1.14 IgG/mL Day 6
Ave = 3.4 x 104 ng/mL
Ave = 8.57 x 104 ng/mL
Ave = 1.72 x 105 ng/mL
Ave = 2.18 x 104 ng/mL
Ave = 4.85 x 104 ng/mL
Ave = 9.89 x 104 ng/mL
*
miR-23 spg panelmiR-NC spg panel
Clone
0.64-fold
0.56-fold
0.57-fold
Day 2
Day 4
Day 6
IgG
ng
/mL
(x
10
4)
IgG
ng
/mL
(x
10
5)
IgG
ng
/mL
(x
10
5)
0
0
0
Day 2
Day 4
Day 6
0.5
1
1.5
2
2.5
3
3.5
0.5
1
1.5
2
2.5
1
2
3
4
5
6
7
8
9
2
2
220
20
20 15
15
15 6
6
6 9
9
917
17
1714
14
1423
23
23
209
Figure 3.51: CHO-1.14 (IgG) miR-23b sponge clonal panel – Specific productivity
Figure 3.51: Specific productivity for CHO clones engineered with a sponge specific for miR-23
versus a control non-specific sponge. Average specific productivity is represented a broken red line.
Clones selected for scale up are highlighted in an orange box. Statistical analysis was calculated using
a standard student t-test (p ≤ 0.05* and p ≤ 0.01**).
As improved viability was observed to be potentially mediated through miR-23 depletion,
all clones selected from the miR-23 panel demonstrated a viability above 80%.
Additionally, although enhanced cellular proliferation was observed within the mixed
pools to the detriment of productivity, a combination of both proliferation and
productivity was considered when selecting clones. For example, of the 5 best performing
miR-23 sponge clones (clone 5, 7, 8, 11 and 24), when viable cell numbers were assessed
0
20
40
60
80
100
120
140
160
14 8 2 4 6 11 15 20 3 21 13 22 1 17 16 9
IgG
pg
/cel
l/d
ay
14 12 2 10 11 6 13 17 16 5 4 7 21 8 22 18 9 20 15 24 23 3
0
10
20
30
40
50
60
70
80
90
14 2 8 4 15 6 20 3 11 21 13 22 17 1 9 16
IgG
pg
/cel
l/d
ay
12 2 14 6 10 11 13 16 17 5 4 22 21 7 8 18 20 15 9 3 23 24
Day 2
Day 4
0
20
40
60
80
100
120
22 2 8 4 14 20 15 21 6 11 3 13 17 1 9 16
IgG
pg
/cel
l/d
ay
Day 6
21 17 2 23 14 13 4 6 9 11 22 12 15 10 16 18 20 3 8 5 7 24
miR-23 spg panelmiR-NC spg panel
Clone
Ave Qp = 59.65
pg/cell/day
Ave Qp = 34.33
pg/cell/day
Ave Qp = 15.94
pg/cell/day
Ave Qp = 13.65
pg/cell/day
Ave Qp = 31.88
pg/cell/day
Ave Qp = 31.57
pg/cell/day
0.57-fold
0.50-fold
0.42-fold
*
**
210
(Fig. 3.48Day 6), only 1 clone (clone 11) remained in the topmost performing group of
clones when volumetric IgG secretion was assessed, with the remaining 4 clones
producing below detectable limits. For these reasons clones were chosen based on day 6
volumetric IgG values (miR-23 sponge clones 6, 9, 14, 17 and 23) as particular clones
(clone 9 and 23) demonstrated high secreted IgG levels at this time point while producing
below the ELISA sensitivity on days 2 and 4. Clonal selection from the miR-NC sponge
panel (clones 2, 15 and 20) was based on similar criteria.
miR-23b* sponge clonal panel - Batch
Next, the influence of miR-23b* depletion was assessed in clones derived from the stable
IgG-secreting CHO-1.14 mixed pool. These were observed previously to have an
enhanced proliferation capacity of 20-50% at the expense of productivity. Furthermore,
cell viability late into the culture process was increased by 10-15% when compared to
controls.
An increase in growth of ~50% (p ≤ 0.001) was observed on day 2 of culture when
compared to control sponge panels (Fig. 3.52Day 2) with a ~30% (p ≤ 0.01) improved
maximal cell density. However, it was noted that particular NC-sponge clones (clone 20,
21 and 16) reached similar cell numbers at this time point (Fig. 3.52Day 4). 20 of 22 (90%)
of miR-23b* sponge clones exceeded the average growth performance of the control
panel. Furthermore, an average 50% (p ≤ 0.01) increase in cell density was maintained at
day 6 of culture potentially resulting from an extended viability (Fig. 3.52Day 6).
An average additional 25% (p ≤ 0.001) viability was observed when compared to the
average of the control panel (87.9% versus 63.4%, Fig. 3.53), a phenotype previously
seen to a lesser extent in mixed pools.
211
Figure 3.52: CHO-1.14 (IgG) miR-23b* sponge clonal panel – Cell numbers/mL
Figure 3.52: The viable cell numbers/mL for clonal miR-23b* and miR-NC sponge populations
derived from their predecessor mixed pools. Individual clonal performance is represented with the
average performance across the panel being shown as a broken red line. Clones selected for scale up
are highlighted in an orange box. Statistical significance was calculated using a standard student t-
test (p ≤ 0.01** and ≤ 0.001***).
0.00E+00
2.00E+05
4.00E+05
6.00E+05
8.00E+05
1.00E+06
1.20E+06
1.40E+06
1.60E+06
20 16 9 3 21 4 13 1 6 17 14 2 11 15 8 22
Cell
Nu
mb
ers/
mL
Clone
miR-NC sponge panel CHO-1.14
0.00E+00
2.00E+05
4.00E+05
6.00E+05
8.00E+05
1.00E+06
1.20E+06
1.40E+06
1.60E+06
21 11 15 12 19 18 14 22 17 20 16 7 10 24 6 9 5 23 2 8 3 1
Cell
Nu
mb
ers/
mL
Clone
miR-23b* sponge panel CHO-1.14
0.00E+00
5.00E+05
1.00E+06
1.50E+06
2.00E+06
2.50E+06
3.00E+06
3.50E+06
4.00E+06
20 21 16 9 3 15 13 6 14 17 4 1 11 2 8 22
Cell
Nu
mb
ers/
mL
Clone
miR-NC sponge panel CHO-1.14 Day 4
0.00E+00
5.00E+05
1.00E+06
1.50E+06
2.00E+06
2.50E+06
3.00E+06
3.50E+06
4.00E+06
21 22 17 7 18 19 20 8 12 24 11 6 14 9 16 2 5 15 3 1 23 10
Cell
Nu
mb
ers/
mL
Clone
miR-23b* sponge panel CHO-1.14 Day 4
0.00E+00
5.00E+05
1.00E+06
1.50E+06
2.00E+06
2.50E+06
3.00E+06
3.50E+06
4.00E+06
4.50E+06
24 22 8 20 19 6 7 9 18 21 17 3 16 23 12 2 1 11 15 14 5 10
Cell
Nu
mb
ers/
mL
Clone
miR-23b* sponge panel CHO-1.14 Day 4
0.00E+00
5.00E+05
1.00E+06
1.50E+06
2.00E+06
2.50E+06
3.00E+06
3.50E+06
4.00E+06
4.50E+06
16 9 15 3 6 14 4 21 13 2 8 20 1 11 17 22
Cell
Nu
mb
ers/
mL
Clone
miR-NC sponge panel CHO-1.14 Day 6
Clone
miR-NC spg panel miR-23b* spg panel
Ave = 6.08 x 105 Cells/mL
Ave = 1.97 x 106 Cells/mL
Ave = 2.05 x 106 Cells/mL Ave = 3.04 x 106 Cells/mL
Ave = 2.6 x 106 Cells/mL
Ave = 9.18 x 105 Cells/mL
***
**
**
Day 2
Day 4
Day 6
Cel
l N
um
ber
s/m
L (
x1
05)
Cel
l N
um
ber
s/m
L (
x1
06)
Cel
l N
um
ber
s/m
L (
x1
06)
Day 2
Day 4
Day 6
0
0
0
2
4
6
8
10
12
14
16
0.5
1.5
2.5
3.5
1
2
3
4
0.5
1.5
2.5
3.5
4.5
1
2
3
4
2
2
2 20
20
20 15
15
15
15
15
156
6
622
22
22
212
Figure 3.53: CHO-1.14 (IgG) miR-23b* sponge clonal panel – Cell viability %
Figure 3.53: Cellular viability status of both miR-23b* and miR-NC rCHO-1.14 clones late in culture.
Average viability across each clonal panel is represented by a broken red line. Clones that were
ultimately selected for scale up are highlighted in orange boxes. Statistical significance is was
calculated using a standard student t-test (p ≤ 0.001***).
0
10
20
30
40
50
60
70
80
90
100
3 13 2 9 15 16 6 1 11 4 8 14 17 21 20 22
Via
bil
ity
%
Clone
miR-NC sponge CHO-1.14 Day 6 Via
0
10
20
30
40
50
60
70
80
90
100
8 9 1 12 2 3 15 16 23 14 6 22 17 5 11 24 18 7 10 19 21 20
Via
bil
ity
%
Clone
miR-23b* sponge CHO-1.14 Day 6 viability
Clone
miR-NC spg panel miR-23b* spg panel
Ave = 63.4 % Ave = 87.9 %***
213
Figure 3.54: CHO-1.14 (IgG) miR-23b* sponge clonal panel – Volumetric
productivity
Figure 3.54: Volumetric IgG secreted for each clonal in both miR-23b* and miR-NC sponge panels
is represented above with the average performance across the panels shown at a broken red line.
Clones ultimately selected for scale up are highlighted in orange boxes. Statistical significance was
determined using a standard student t-test (p ≤ 0.05*).
Similarly to the miR-23 sponge panel, a significant number of miR-23b* clones
completely lost IgG production but were included in statistical analysis. An average
reduction of 30-45% was observed in volumetric IgG across all time points (Fig. 3.54).
Cell specific productivity was significantly reduced by ~50% (p ≤ 0.05) on day 2 and
~60-65% on day 4 and 6 (p ≤ 0.01), respectively (Fig. 3.55).
0.00E+00
1.00E+04
2.00E+04
3.00E+04
4.00E+04
5.00E+04
6.00E+04
7.00E+04
8.00E+04
9.00E+04
2 14 8 4 20 3 6 11 15 21 13 1 17 16 22 9
IgG
ng
/mL
Clone
miR-NC sponge CHO-1.14 IgG/mL Day 2
0.00E+00
1.00E+04
2.00E+04
3.00E+04
4.00E+04
5.00E+04
6.00E+04
7.00E+04
8.00E+04
9.00E+04
21 23 15 18 11 1 14 9 6 3 22 2 16 12 19 10 24 17 7 8 5 20
IgG
ng
/mL
Clone
miR-23b* sponge CHO-1.14 IgG/mL Day 2
0.00E+00
5.00E+04
1.00E+05
1.50E+05
2.00E+05
2.50E+05
15 23 18 21 1 9 3 6 14 22 2 5 16 12 24 10 19 7 20 17 8 11
IgG
ng
/mL
Clone
miR-23b* sponge CHO-1.14 IgG/mL Day 4
0.00E+00
5.00E+04
1.00E+05
1.50E+05
2.00E+05
2.50E+05
14 20 2 8 3 4 15 6 21 11 13 17 9 1 16 22
IgG
ng
/mL
Clone
miR-NC sponge CHO-1.14 IgG/mL Day 4
0.00E+00
5.00E+04
1.00E+05
1.50E+05
2.00E+05
2.50E+05
3.00E+05
3.50E+05
20 2 14 21 15 8 4 3 6 13 11 22 17 16 9 1
IgG
ng
/mL
Clone
miR-NC sponge CHO-1.14 IgG/mL Day 6
0.00E+00
5.00E+04
1.00E+05
1.50E+05
2.00E+05
2.50E+05
3.00E+05
3.50E+05
15 12 1 11 14 6 10 17 18 23 22 5 3 7 20 8 19 24 21 2 9 16
IgG
ng
/mL
Clone
miR-23b* sponge CHO-1.14 IgG/mL Day 6
Ave = 3.4 x 104 ng/mL
Ave = 8.57 x 104 ng/mL
Ave = 1.72 x 105 ng/mL
Ave = 2.35 x 104 ng/mL
Ave = 4.69 x 104 ng/mL
Ave = 9.51 x 105 ng/mL
*
0.69-fold
0.54-fold
0.55-fold
Clone
miR-NC spg panel miR-23b* spg panel
Day 2
Day 4
Day 6
IgG
ng
/mL
(x
10
5)
IgG
ng
/mL
(x
10
5)
IgG
ng
/mL
(x
10
4)
Day 2
Day 4
Day 6
0
0
0
0.5
1.5
2.5
1
2
0.5
1.5
2.5
3.5
1
2
3
4
5
6
7
8
9
20
20
20 2
2
2 15
15
15 15
15
15 6
6
6 22
22
22
214
Figure 3.55: CHO-1.14 (IgG) miR-23b* sponge clonal panel – Specific productivity
Figure 3.55: Specific productivity for each clone from both miR-23b* and miR-NC sponge panels are
shown above with the average performance represented as a broken red line. Statistics were
calculated using a standard student t-test (p ≤ 0.05* and ≤ 0.01**) with each clone chosen for scale
up highlighted in an orange box.
The significant number of miR-23b* depleted clones that exhibited a reduced IgG titre
below the detectable limits of the ELISA is potentially due to the inherent reduction in
productivity due to the enhanced growth phenotype as previously observed for mixed
populations in CHO-1.14 and additionally conserved in CHO-SEAP stables (not as
dramatic a growth impact). Clones ultimately selected for 5 mL scale up were based on
their viability and cell growth attributes while also maintaining a production capacity.
Although clonal selection proved difficult, it was anticipated that if miR-23b* depletion
1 15 12 14 10 11 6 17 23 5 18 3 22 7 20 8 19 2 9 16 24 21
23 15 1 18 21 3 9 6 14 2 5 10 22 16 8 24 20 12 7 19 17 11
23 21 15 1 18 11 9 14 6 3 22 2 8 16 24 10 5 7 17 12 19 20
0
20
40
60
80
100
120
140
160
14 8 2 4 6 11 15 20 3 21 13 22 1 17 16 9
IgG
pg
/cel
l/d
ay
0
10
20
30
40
50
60
70
80
90
14 2 8 4 15 6 20 3 11 21 13 22 17 1 9 16
IgG
pg
/cel
l/d
ay
Day 2
Day 4
0
20
40
60
80
100
120
22 2 8 4 14 20 15 21 6 11 3 13 17 1 9 16
IgG
pg
/cel
l/d
ay
Day 6
miR-NC spg panel
Clone
Ave Qp = 59.65
pg/cell/day
Ave Qp = 31.35
pg/cell/day
Ave Qp = 12.60
pg/cell/day
Ave Qp = 10.85
pg/cell/day
Ave Qp = 31.88
pg/cell/day
Ave Qp = 31.57
pg/cell/day
0.52-fold
0.39-fold
0.34-fold
**
**
*
2
2
2 20
20
2015
15
15 15
15
15 22
22
226
6
6
215
mediated apoptosis resistance in selected clones, their inherent reduced specific
productivity would be overcome by a potential prolonged production phase, potentially
achieving higher volumetric titres. Clones 6, 15 and 22 were selected from the miR-23b*
sponge panel and clones 2, 15 and 20 were selected from the miR-NC sponge panel, as
before for miR-23 sponge comparison.
3.3.6 Investigation of miR-23 and -23b* depletion on selected clones in 5mL Batch,
Fed-Batch and Fed-batch temperature shift
This next section characterises the performance of clonal behaviour for both stable miR-
23 and miR-23b* sponge CHO-SEAPs over a serious of commonly implemented
bioprocess regimes including batch, fed-batch and temperature shift fed-batch.
3.3.6.1 miR-23 and miR-23b* sponge CHO-SEAP clones
miR-23 sponge clones - Batch
Clones selected from the clonal panel screen (Clones 2, 4, 6 and 23) were seeded at 2 x
105 cells/mL in a 5 mL working volume in a 50 mL vented spin tube (37 ˚C at 170 rpm).
A slight variation in maximal cell density was evident between all clones to within a range
of 1 x 106 /mL, potentially attributed to clonal variation (Fig. 3.56 A). A range of
viabilities was observed with a single control clone remaining above 80%, potentially an
innate clonal benefit (Fig. 3.56 B).
Productivity was represented as a range with each test clone being compared to the best
and worst performing control clone. Statistical analysis was calculated for each test clone
against the pooled values for each control clone. The best performing clones were miR-
23 spg 4 and 6, both demonstrating similar SEAP activity increases of 2-6-fold (p ≤ 0.001)
while exhibiting a 2-5-fold enhanced normalised productivity as a result of a slightly
higher cell density (Fig. 3.56 C and D). The lowest performing miR-23 sponge clone
(miR-23 spg 2) still maintained a higher SEAP activity when compared to the best
performing control clone ranging from 1.25-3-fold (p ≤ 0.001).
216
A note to make at this point would be the apparent dramatic increase in normalised
productivity of the control clone, miR-NC spg 20. When cultured in 5 mL volumes, this
clone had a propensity to clump therefore exhibiting an apparent reduced cell density thus
resulting in an enhanced normalised productivity calculation. Subsequent bioprocess
characterisation was carried out with this clone omitted. Additionally, miR-23 sponge
clone 4 and 6 were retained as they both exhibited similar characteristics.
Figure 3.56: miR-23 sponge clones 2, 4, 6 and 23 under 5 mL batch conditions
Figure 3.56: CHO-SEAP stable clones 2, 4, 6 and 23 stably depleted of miR-23 grown under batch
condition in a 5 mL volume seeded at 2 x 105 cell/mL are represented above in different shades of red
while miR-NC clones 2, 11 and 20 are shown in various shades of black (n = 3). Clone performance
is reported in text as a range compared to the best and worst control clone while statistical
significance was calculated based on the combined values for all control clones (p ≤ 0.001***,
standard student t-test).
It is evident that miR-23 depleted clones selected from the mixed pool exhibited a
maximum 2-fold increase in volumetric SEAP activity when compared to their mixed
population predecessors under batch conditions (Fig. 3.57), highlighting the benefits of
clonal isolation.
0
200
400
600
800
1000
1200
0 2 4 6
SE
AP
Un
its/
mL
Day
miR-23b spg clones SEAP
miR-NC spg 2
miR-NC spg 11
miR-NC spg 20
miR-23b spg 2
miR-23b spg 4
miR-23b spg 6
miR-23b spg 23
0
200
400
600
800
1000
1200
1 2 3 4 5 6 7
SE
AP
Un
its/
mL
Day
0
1
2
3
4
5
6
1 2 3 4 5 6 7
Cell
s/m
L (
x1
06)
Day
0
0.5
1
1.5
2
2.5
3
3.5
4
4.5
5
1 2 3 4 5 6 7
SE
AP
Un
its/
cell
(x1
0-4
)
Day
0
20
40
60
80
100
120
1 2 3 4 5 6 7
Via
bil
ity
%
Day
A B
C D
***
***
***
217
Another interesting observation was that a similar SEAP activity was observed in miR-
23 sponge clones cultured in 1 mL and 5 mL batch culture despite reduced maximal cell
densities achieved in 1 mL culture. This SEAP activity was attributed to an increased
normalised productivity (~2-fold) for those clones that reached a lower cell density.
Figure 3.57: Volumetric SEAP performance of miR-23 sponge clones versus miR-23
sponge mixed pools
Figure 3.57: Various performing clones that make up a mixed population as a result of the nature of
plasmid integration and expression upon transgenic CHO cell line generation. The SEAP activity is
compared between the predecessor miR-23 sponge mixed CHO cell population and the subsequently
isolated clones.
miR-23 sponge clones – Fed-batch
Next we assessed whether a fed-batch process would further enhance cell density of these
high SEAP producing clones potentially improving productivity. miR-23 sponge clones
4 and 6 were evaluated as they both demonstrated similar attributes as well as miR-NC
spg 2 and 11. Clonal variation was once again evident upon the addition of a feed with a
range of maximal cell densities being observed (Fig. 3.58 A).
0
200
400
600
800
1000
1200
1 2 3 4 5 6 7
SE
AP
Un
its/
mL
Day
0
200
400
600
800
1000
1200
1 3 5 7
SE
AP
Un
its/
mL
Day
miR-23b spg clones SEAP
miR-23b spg 2
miR-23b spg 4
miR-23b spg 6
miR-23b spg 23
miR-23b spg MP
218
Figure 3.58: miR-23 sponge clone 4 and 6 under 5 mL Fed-Batch conditions
Figure 3.58: An 8 day fed-batch culture with a working volume of 5 mL (+ initial feed of 15% v/v,
represented with a broken orange line) seeded at 2 x 105 cell/mL with subsequent feeding of 10%
(v/v) every 3 days. A) Cell numbers/mL, B) Cell viability with 80% viability threshold represented as
a broken red line, C) Volumetric SEAP activity/mL with statistical significance calculated based on
each test clone versus all three combined control clones (p ≤ 0.001, student t-test) and D) specific
SEAP activity per cell, n = 2.
Both miR-23 sponge clones 4 and 6 exhibited an enhanced viability of ~20% (p ≤ 0.01)
on days 6, 7 and 8 of culture when compared to control clones while also maintaining a
prolonged viability above the 80% threshold for an additional 48 h (Fig. 3.58 B). Both
miR-23 depleted clones exhibited a 2-6-fold (p ≤ 0.001) increase in SEAP activity when
compared to controls (Fig. 3.58 C). Given the various growth phenotypes observed,
normalised productivity was improved but not proportional to volumetric productivity
(Fig. 3.58 D). It was observed that that maximum SEAP produced was reduced in Fed-
batch for miR-23 depleted clones when compared to batch despite the improved cell
densities, possibly due to a compromise in normalised SEAP production.
0.00
2.00
4.00
6.00
8.00
10.00
12.00
0 2 4 6 8
Cell
s/m
L (
x1
06)
Day
0.00
2.00
4.00
6.00
8.00
10.00
12.00
0 2 4 6 8S
EA
P U
nit
s/cell
(x1
0-5
)
Day
**
****
A
25
35
45
55
65
75
85
95
0 1 2 3 4 5 6 7 8
Via
bil
ity
%
Day
Viabile cell density
0
100
200
300
400
500
600
700
800
0 2 4 6 8
SE
AP
Un
its/
mL
Day
Volumetric Productivity
miR-NC spg 2
miR-NC spg 11
miR-23b spg 4
miR-23b spg 6
0
100
200
300
400
500
600
700
800
0 2 4 6 8
SE
AP
Un
its/
mL
Day
Volumetric Productivity
***
***
******
***
***
***
***
B
DC
219
miR-23 sponge clones – Fed-batch with Temperature shift
Another commonly implemented process utilised in industry is hypothermic growth
conditions (Kumar, Gammell & Clynes 2007). By reducing the culture temperature from
37 ˚C to ~31 ˚C, specific CHO cell productivity can be enhanced through arrest in the G1-
phase of the cell cycle, the most transcriptionally active phase.
With the observation that fed-batch boosted both growth and maximum cell density of
our miR-23 sponge CHO-SEAP clones with the addition of imparting an extended viable
production phase above 80% for a further 48 h, we sought to overcome the reduced SEAP
activity when compared to batch culture by inducing a temperature shift thus potentially
enhancing specific productivity and prolonging culture longevity.
Fed-batch was carried out as previously described with a temperature shift to 31 ˚C after
72 h of culture. Both miR-23 and miR-NC sponge clones performed similar over the first
three days of fed-batch culture reaching cell densities of ~ 6 x 106 cells/mL prior to
temperature shift, as observed previously. Cellular proliferation was observed to be halted
upon shift to 31 ˚C three days into culture for all clones except miR-NC spg 2 which
climbed a further 2 x 106 cell/mL at its maximum (Fig. 3.59 A). A gradual decrease in
cell density, forming the induced production phase, can be observed upon TS which was
due to a combination of both cell death and cellular clumping. Another interesting
observation was that miR-23 sponge clones appeared to be sensitive to TS demonstrating
a negative response to hypothermic culture conditions dropping below the 80% viability
threshold on day 5 (Fig. 3.59 B) when compared to day 7 in the absence of TS (Fig. 3.58
B). However, a viable production phase of ~60% was maintained as far as day 13 of
culture. Furthermore, miR-NC clones declined faster in viability at very late culture time
points with a significant separation observed of ~10-20% (p ≤ 0.05) (Fig. 3.59 B).
Enhanced SEAP productivity of 2-4-fold (Volumetric, p ≤ 0.001) and 2-5-fold
(Normalised) respectively, was observed for miR-23 depleted clones (Fig. 3.59 C and D).
220
Figure 3.59: miR-23 sponge clones under Fed-batch temperature shift
Figure 3.59: miR-23 sponge clones 4 and 6 were subjected to fed-batch (Orange broken line)
temperature shift (broken blue line) as a means to exploit the increased cell densities accumulated in
early culture while subsequently potentially improving specific productivity and reducing the onset
of apoptosis. Nutrient feeds were as described in the case of a fed-batch. Phenotypes assessed were
A) Viable cell numbers/mL, B) Cellular viability with an 80% viability threshold represented as a
broken red line, C) Volumetric SEAP activity and D) specific productivity. Statistics were calculated
using a standard student t-test (p ≤ 0.05, p ≤ 0.001***) and an n = 2.
0
1
2
3
4
5
6
7
0 2 4 6 8 10 12 14
SE
AP
Un
its/
cell
(x1
0-4
)
Day
0
20
40
60
80
100
120
0 1 2 3 4 5 6 7 8 9 10 11 12 13
Via
bil
ity
%
Day
miR-23b spg clones
0
200
400
600
800
1000
1200
0 2 4 6 8 10 12 14
SE
AP
Un
its/
mL
Day
Fed-BatchTS
miR-NC spg 2
miR-NC spg 11
miR-23b spg 4
miR-23b spg 6
0
200
400
600
800
1000
1200
0 2 4 6 8 10 12 14
SE
AP
Un
its/
mL
Day
Fed-BatchTS
***
***
*** ******
******
*** *** ***
*** ***
D
C
B
A
*
0.00
2.00
4.00
6.00
8.00
10.00
12.00
0 2 4 6 8 10 12 14
Cel
l N
um
ber
s/m
L (
x1
06)
Day
221
Normalised productivity was observed to be increased in fed-batch with temperature shift
when compared to fed-batch only for miR-23 sponge clones, with normalised
productivity spikes in late culture potentially attributed to clumping (Fig. 3.60).
Figure 3.60: miR-23 sponge 4 and 6 clones specific productivity under Fed-batch
and fed-batch temperature shift
Figure 3.60: The influence of a fed-batch versus a fed-batch with temperature shift on specific SEAP
productivity is compared above for clone 4 and 6 that have been engineered to stably deplete miR-
23. Nutrient feeds are as previously described (15% v/v on Day 0 and 10% v/v every 72 hs) and
represented as a broken orange line. Temperature shift (TS) to 31˚C is denoted on day 3 as a broken
blue line.
miR-23b* sponge clones - Batch
miR-23b* sponge SEAP clones (clone 3,8 and 18) were selected based on their growth
and productivity phenotype when evaluated under 1 mL suspension.. A range of growth
phenotypes were observed when clones were cultured under 5 mL batch conditions. One
clone in particular, miR-23b* spg 18, demonstrated an enhanced growth phenotype on
day 4 of culture achieving an improved maximum cell concentration of 42% (p ≤ 0.01)
when compared to the highest performing control clone (Fig. 3.61 A). Furthermore, this
clone maintained a high viable cell density of 91.1% (p ≤ 0.05) late in culture when
compared to controls (66.4, 81.6 and 27.8) (Fig. 3.61 B). When normalised productivity
was compared to miR-NC spg 11, the highest producing control clone, all miR-23b* spg
0
0.00005
0.0001
0.00015
0.0002
0.00025
0.0003
0 2 4 6 8 10
SE
AP
Un
its/
cell
Day
miR-23b spg 4 TS
miR-23b spg 6 TS
miR-23b spg 4
miR-23b spg 6
0
5
10
15
20
25
30
0 2 4 6 8 10
SE
AP
Un
its/
cell
(x1
0-5
)
Day
222
clones demonstrated a reduced output across all time points ranging from 75-35% (Fig.
3.61 D). However, miR-23b* spg 18 produced an equivalent volumetric SEAP yield on
day 6 despite this observed reduced normalised SEAP activity (Fig. 3.61 C). This
observation was potentially due to the enhanced maximal cell density achieved coupled
with the prolonged viability thus mediating a prolonged viable production phase.
With the observation of this highly attractive phenotype, miR-23b* spg clone 18 was
carried on for further validation using various culture conditions.
Figure 3.61: miR-23b* sponge CHO-SEAP clones under batch conditions
Figure 3.61: miR-23b* sponge CHO-SEAP clones cultured under batch conditions are represented
above for all bioprocess phenotypes including A) Cell numbers/mL, B) Cellular viability, C)
Volumetric SEAP activity/mL and D) Specific SEAP activity/cell. Statistical significance was
calculated using the standard student t-test (p ≤ 0.01**) and an n = 3.
miR-23b* sponge clone 18 – Fed-batch
A similar phenotype was observed when miR-23b* sponge clone 18 was cultured in the
presence of a feed as had been observed when under batch conditions. The addition of a
feed did not boost maximal cell densities achieved by the miR-23b* spg 18 clone as its
0
1
2
3
4
5
6
7
8
1 2 3 4 5 6 7
Cel
ls/m
L (
x1
06)
Day
0
20
40
60
80
100
120
1 2 3 4 5 6 7
Via
bili
ty %
Clone
Viability
miR-NC 2
miR-NC 11
miR-NC 20
miR-23b* 3
miR-23b* 8
miR-23b* 18
0
50
100
150
200
250
300
350
400
450
1 2 3 4 5 6 7
SE
AP
Un
its/
mL
Day
miR-23b* spg clones SEAP
**
**
0
20
40
60
80
100
120
1 2 3 4 5 6 7
Via
bili
ty %
Day
Viability
*
A B
C D
0
1
2
3
4
5
6
7
8
9
10
1 2 3 4 5 6 7
SE
AP
Un
its/
cell
(x1
0-4
)
Day
223
growth profile was previously observed to be high under batch conditions. This clone
demonstrated an intermediate growth profile when compared to both miR-NC 2 and miR-
NC 11 (Fig. 3.62 A). An extended viability was observed for clone 18 with a further 48
hs of culture above 80% viability (p ≤ 0.001) when compared to control clones (Fig. 3.62
B). The benefit of this extended viable production phase was again reflected in the gradual
increase in volumetric SEAP activity of miR-23b* spg 18 producing ~10% more than the
highest producing control clone (Fig. 3.62 C) despite a reduced normalised productivity,
75-30% (Fig. 3.62 D).
Figure 3.62: miR-23b* sponge clone 18 CHO-SEAP under fed-batch conditions
Figure 3.62: Fed-batch culture results for miR-23b* spg 18 versus two top performing control sponge
clones (miR-NC spg 2 and 11). Cells were seeded at 2 x 105 cells/mL with an initial feed of 15% (v/v)
and an additional subsequent feed of 10% (v/v) every 72 h represented as a broken orange line.
Graphs demonstrate A) Cell numbers/mL, B) Cell viability with an 80% viability threshold shown
as a broken red line, C) Volumetric SEAP activity/mL and D) Normalised productivity. Statistical
significance was calculated using the standard student t-test (p ≤ 0.001***), n = 2.
0.00
2.00
4.00
6.00
8.00
10.00
12.00
0 2 4 6 8
SE
AP
Un
its/
cell
(x1
0-5
)
Day
0.00
1.00
2.00
3.00
4.00
5.00
6.00
7.00
8.00
9.00
0 1 2 3 4 5 6 7 8 9
Cel
ls/m
L (
x1
06)
Day
25
35
45
55
65
75
85
95
0 1 2 3 4 5 6 7 8
Via
bil
ity
%
Day
Viability % miR-23b* spg 18
******
***
***
0
20
40
60
80
100
120
1 2 3 4 5 6 7
Via
bili
ty %
Clone
Viability
miR-NC 2
miR-NC 11
miR-NC 20
miR-23b* 3
miR-23b* 8
miR-23b* 18
0
20
40
60
80
100
120
1 2 3 4 5 6 7
Via
bili
ty %
Clone
Viability
miR-NC 2
miR-NC 11
miR-NC 20
miR-23b* 3
miR-23b* 8
miR-23b* 18
0
50
100
150
200
250
0 2 4 6 8
SE
AP
Un
its/
mL
Day
miR-23b* spg 18 Volumetric Titre
A B
C D
224
miR-23b* sponge clone 18 – Fed-batch with Temperature shift
With the observation that this miR-23b* sponge clone exhibited a beneficial viability
phenotype when compared to controls that ultimately improved the production
phenotype, we sought to further enhance the production phase through the utilisation of
hypothermic growth conditions.
miR-23b* spg clone 18 exhibited similar cell densities to controls over the three days
prior to TS with an immediate growth arrest being observed upon transfer to 31 ˚C (Fig.
3.63 A). A prolonged viable production phase was observed in the case of miR-23b*
depletion in clone 18 (p ≤ 0.001-0.05 at various time points) while remaining above 80%
viability for a further 5 days when compared to controls (Fig. 3.63 B). Volumetric yield
was boosted by 20-80% (p ≤ 0.001) over various time points (Fig. 3.63 C). Normalised
productivity was observed to be lower than controls at the early stages of the culture
process but appeared to be gradually improved later into culture due to the gradual
reduction in cell numbers and the increase in SEAP activity (Fig. 3.63 D).
From progressing particular CHO clones generated from both the miR-23 sponge and
miR-23b* panels through various culture processes, it was evident that both miRNA
engineering approaches endowed a beneficial CHO cell phenotype although miR-23
depletion was more consistent than miR-23b*, which allowed for the isolation of a single
clone (clone 18) with an enhanced viability performance. The observed prolonged
viability suited this miR-23b* sponge clone to the bioprocess of fed-batch with
temperature shift. miR-23 sponge clones exhibited an attractive phenotype suitable to
several production platforms (batch, fed-batch and fed-batch with TS) and for this reason
was pursued for further validation.
225
Figure 3.63: miR-23b* sponge clone 18 CHO-SEAP under batch conditions
Figure 3.63: Fed-batch (Orange broken line) culture with an additional temperature shift to 31 ˚C
(Broken blue line) for the CHO-SEAP clone 18 stably depleted of miR-23b* and control sponge
clones. A) Cell numbers/mL, B) cell viability with an 80% viability threshold represented as a broke
red line, C) Volumetric SEAP activity/mL and D) Specific SEAP activity. Statistical analysis was
calculated using the standard student t-test (p ≤ 0.05*, ≤ 0.01** and ≤ 0.001***), n = 2.
0
0.5
1
1.5
2
2.5
3
3.5
4
0 2 4 6 8 10 12 14
SE
AP
Un
its/
cell
(x1
0-4
)
Day
0
100
200
300
400
500
600
700
800
0 2 4 6 8 10 12 14
SE
AP
Un
its/
mL
Day
0
10
20
30
40
50
60
70
80
90
100
0 1 2 3 4 5 6 7 8 9 10 11 12 13
Via
bil
ity
%
Day
miR-23b* spg clone 18
*** *** * * * ** ** **
***
**
***
*** ***
***
***
0.00E+00
2.00E+06
4.00E+06
6.00E+06
8.00E+06
1.00E+07
1.20E+07
0 2 4 6 8 10 12 14
Cel
l Nu
mb
ers/
mL
Day
miR-23b* spg clones Fed-BatchTS
0
20
40
60
80
100
120
1 2 3 4 5 6 7
Via
bili
ty %
Clone
Viability
miR-NC 2
miR-NC 11
miR-NC 20
miR-23b* 3
miR-23b* 8
miR-23b* 18
0
20
40
60
80
100
120
1 2 3 4 5 6 7
Via
bili
ty %
Clone
Viability
miR-NC 2
miR-NC 11
miR-NC 20
miR-23b* 3
miR-23b* 8
miR-23b* 18
A
B
C
D
***
226
3.3.6.2 miR-23 sponge CHO-1.14 (IgG) clones
miR-23 sponge clones - Batch
As reported previously, miR-23 depletion in the IgG-secreting CHO-1.14 cells was
different from what was observed in SEAP cells. Growth and viability were enhanced in
the case of CHO-1.14 to the detriment of productivity. Furthermore, it was observed
within the clonal panel study that viability was improved by ~20%. Although, volumetric
and specific productivity was observed to be reduced, miR-23 sponge CHO-1.14 clones
were selected based on IgG levels, i.e. attempting to isolate the best producers while
considering both growth and viability traits. From the clones chosen (miR-23 spg 6, 9,
14, 17 and 23), various bioprocess conditions were tested.
Enhanced growth of ~1.5-2-fold (p ≤ 0.001) on days 2 and 4 of batch culture was observed
for all miR-23 sponge clones when compared to all three control clones (Fig. 3.64 A).
One control clone (miR-NC spg 20) demonstrated a comparable viability (above 80%) to
that of miR-23 depleted clones unlike its counterparts whose viability on day 6 of culture
were 54% and 22% (Fig. 3.64 B). Although all 5 miR-23 depleted clones demonstrated
higher viability than the two lowest performing control clones, only two retained
enhanced viability over the best performing control clone. It is still evident from the
number of clones selected that miR-23 depletion does have a positive and beneficial
impact on CHO cell growth.
227
Figure 3.64: miR-23 depleted CHO-1.14 clones under batch conditions
Figure 3.64: CHO-1.14 clones stably engineered to be depleted of miR-23 under batch culture
conditions. A) Cell numbers/mL, B) Cellular viability, C) Volumetric productivity (ng/mL) and D)
Specific productivity. Statistical analysis was determined using a standard student t-test and was
based on a comparison between each individual test clone and all control clones (p ≤ 0.001***), n =
3.
All miR-23 sponge clones demonstrated a ~40-60% reduction in volumetric IgG when
compared to the best performing control clone, miR-NC sponge clone 2 (Fig. 3.64 C).
When compared to the lower producing controls, there was a less dramatic reduction in
volumetric IgG titre observed of ~20% as far as +20% in some instances, for all miR-23
sponge clones. However, all miR-23 depleted clones demonstrated a reduction in specific
productivity ranging from 25-90% (Fig. 3.64 D) when compared to control clones. It is
evident that from batch cultivation of a selected panel of miR-23 sponge clones in a
recombinant IgG-secreting CHO cell line exhibit an enhanced growth phenotype at the
expense of specific productivity.
0.00
0.50
1.00
1.50
2.00
2.50
3.00
3.50
4.00
4.50
0 1 2 3 4 5 6 7
Cel
l N
um
ber
s/m
L (
x1
06)
Day
0.00E+00
5.00E+05
1.00E+06
1.50E+06
2.00E+06
2.50E+06
3.00E+06
3.50E+06
4.00E+06
4.50E+06
0 2 4 6 8
Cell
Nu
mb
ers/
mL
Day
miR-23 spgCHO-1.14 clones
miR-NC spg 2
miR-NC spg 15
miR-NC spg 20
miR-23 spg 6
miR-23 spg 9
miR-23 spg 14
miR-23 spg 17
miR-23 spg 23
0
20
40
60
80
100
120
0 1 2 3 4 5 6 7
Via
bil
ity
%
Day
miR-23 spg CHO-1.14 viability A B
C D
***
***
0
2
4
6
8
10
12
14
0 1 2 3 4 5 6 7
IgG
ng
/mL
(x1
04)
Day
0
10
20
30
40
50
60
0 1 2 3 4 5 6 7
IgG
pg
/cel
l/d
ay
Day
228
miR-23 sponge clones – Fed-batch
We sought to attempt to further boost cell numbers with the addition of a feed and
potentially mediate a prolonged production phase to ultimately improve volumetric titres
considering the reduced specific productivity. One immediate observation was that the
addition of a feed did not appear to enhance cellular proliferation past normal levels, a
phenotype which had been previously noted in the case of CHO-SEAP during fed-batch.
miR-23 depleted clones demonstrated an enhanced growth phenotype early into culture
of ~50% on day 1 and 2 while achieving an increased maximal cell density of ~50-70%
on day 3 when compared to control clone 2 and 15 (Fig. 3.65 A). One control clone
however (miR-NC spg 20), as observed in batch culture, demonstrated similar maximal
cell densities and a superior viability profile (Fig. 3.65 B). Although culture under batch
conditions suggested a potential viability benefit, there was no discernible improvement
observed when a feed was added.
Figure 3.65: miR-23 depleted CHO-1.14 clones under fed-batch conditions
Figure 3.65: miR-23 depleted clones secreting IgG under fed-batch conditions with the broken orange
line indicating the addition of a nutrient feed. Phenotypes monitored were A) cell cumbers/mL, B)
Cellular viability with an 80% threshold represented as a broken red line, C) Volumetric IgG titres
and D) Specific productivity (n =2).
0
0.5
1
1.5
2
2.5
0 2 4 6 8 10
IgG
ng
/mL
(x1
05)
Day
-0.5
0
0.5
1
1.5
2
2.5
3
0 2 4 6 8 10
Cel
l N
um
ber
s/m
L (
x1
06)
Day
-5.00E+05
0.00E+00
5.00E+05
1.00E+06
1.50E+06
2.00E+06
2.50E+06
3.00E+06
0 2 4 6 8 10
miR-23b spg clonesCHO-1.14
miR-NC-2
miR-NC-15
miR-NC-20
miR-23b-6
miR-23b-9
miR-23b-23
0
20
40
60
80
100
120
0 2 4 6 8 10
Via
bil
ity
%
Day
miR-23b spg clonesCHO-1.14
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
0 2 4 6 8 10
IgG
ng
/cell
Day
miR-23 spg CHO-1.14 Qp FB
A B
C D
229
The three miR-23 sponge clones selected for fed-batch culture demonstrated the highest
cell densities when grown under batch conditions and demonstrated a potential improved
volumetric IgG titre at various time points when compared to the two lowest producing
control clones. In the case of fed-batch, this improved volumetric productivity was more
potent in the case of miR-23 spg clone 6 and 9 with a maximum increased IgG titre on
day 9 of ~40-80% and 2-2.6-fold, respectively. miR-23 spg clone 23 exhibited the lowest
productivity profile of all clones under investigation while miR-NC spg clone 2 surpassed
all clones when both volumetric and specific productivity were assessed (Fig. 3.65 C and
D). Specific IgG productivity was observed to be enhanced for 2 out of three clones when
fed-batch was carried out possibly due to the observation of no dramatic improvement in
cell growth, potentially resulting in an excess of nutrients available for protein synthesis.
This demonstrates that miR-23 depletion in our IgG-secreting CHO-1.14 cell line
improves cellular growth and maximal cell density with a benefit to viability. These
enhanced phenotypes appear to be at the cost of productivity. However, it is evident that
the process of clonal selection and the inherent heterogeneity present within CHO cell
clones can in itself mediate a beneficial phenotype that can be selected for. The selection
of a larger panel of clones would give a better understanding of the impact that miR-23
depletion has on this CHO-1.14 cell line.
At this point however, the depletion of miR-23 in CHO-SEAP cells was more beneficial
and consistent upon scale up and will be the primary focus for the rest of this thesis. As
both IgG and SEAP secreting CHO cell lines were derived from the same parental cell
line, it is possible that the impact of stable miR-23 depletion is product specific with the
culture format also having an influence. The fact that clones selected throughout the
various phases of scale-up were selected based on their volumetric productivity could
have introduced a bias. However, the initial phenotype observed within mixed pools was
observed to be retained throughout.
230
Section 3.4
Investigation of the molecular
impact of miR-23 depletion in
CHO-SEAP cells
231
3.4.1 miR-23 depletion improves CHO-SEAP specific productivity
miR-23 depleted clones exhibited an improved SEAP productivity of ~3-fold with a slight
variation in cellular growth over various culture conditions. This section focuses on the
elucidation of how the depletion of miR-23 is conferring this advantage on cellular
productivity.
3.4.2 SEAP transcript levels appear elevated in miR-23 sponge clones
As all secreted mammalian proteins are translocated to the lumen of the endoplasmic
reticulum either post- or co-translationally for modification, folded and assembled before
release to the cis-Golgi compartment (Tigges, Fussenegger 2006), they are susceptible to
secretory bottlenecks in the event of high protein turnover. We evaluated the transcript
levels of SEAP within a small panel of selected clones. Clones were cultured under the
same batch conditions as reported previously in section 3.3.6.1 and harvested for RNA
(see section 2.4.6 of Materials and Methods) on day 3 of culture. Semi-quantitative PCR
indicated that SEAP mRNA levels were elevated in all test clones when compared to
control (Fig. 3.66 A).
A more definitive comparison in the levels of SEAP mRNA abundance in our miR-23
engineered clones was carried out using qRT-PCR (see section 2.4.4 of Materials and
Methods). The intermediate SEAP producing control clone (miR-NC spg 11) was chosen
as the calibrator when data was normalised during data analysis. A 2-3-fold increase in
SEAP transcript (p ≤ 0.05 and 0.01) was observed for all miR-23 sponge clones (Fig. 3.66
B). Significance was calculated based on Ct (Cyclic threshold) values for each individual
miR-23 sponge clone versus a combination of all three controls. Although error bars to
appear to cover a significant portion of the overall dataset, the Ct values were quite tight
and resulted in significant statistics when a t-test was carried out. Inclusion of a larger
number of technical replicates would have aided in reducing the level of variation within
samples ultimately contributing to a reduction in error.
These results would suggest that the enhanced SEAP yields observed are potentially as a
result of an elevation in the rate of transcription of the SEAP gene in miR-23 depleted
cells.
232
Figure 3.66: Semi-quantitative/qPCR for SEAP transcript in miR-23 sponge clones
Figure 3.66: A) Semi-quantitative gel-based PCR and quantitative real-time PCR (qPCR) was
carried out on miR-23 depleted CHO-SEAP clones in an attempt to assess the potential difference in
transcript levels. A) Represents the ethidium bromide (EtBr)-stained agarose gel primed for the
SEAP mRNA transcript in all miR-23b clones and control clones. B) Quantitative real-time PCR was
carried out to accurately determine the relative fold change in transcript levels of SEAP. miR-NC
spg 20 was chosen as the clone to calibrate the relative quantity (RQ) of SEAP across all clones while
statistical analysis was determined for each individual test clone by using a standard student t-test
for all Ct values across all three controls, (p ≤ 0.05* and p ≤ 0.01**).
When the mature levels of miR-23b were assessed by qRT-PCR (see section 2.4.8 of
Materials and Methods), there was a 5-fold (p ≤ 0.001***) reduction observed in clones
expressing a miR-23 specific sponge compared to controls (Fig. 3.67). This reduction in
mature miR-23 levels was similar to what was observed in the case of miR-34a.
0
0.5
1
1.5
2
2.5
3
3.5
miR-NC spg 2
miR-NC spg 11
miR-NC spg 20
miR-23b spg 2
miR-23b spg 4
miR-23b spg 6
miR-23b spg 23
RQ
Clone
SEAP mRNA levels
0
0.5
1
1.5
2
2.5
3
3.5
RQ
Clone
SEAP mRNA levels
SEAP mRNA
*
**
**
*
A
B
233
Figure 3.67: Endogenous miR-23b levels in miR-23 sponge clones
Figure 3.67: qRT-PCR was utilised to assess mature levels of endogenous miR-23b in miR-23 sponge
clones when compared to control clones. Significance was calculated for each miR-23 clone across all
3 miR-NC sponge clones using a standard student t-test (p ≤ 0.001***).
3.4.3 SEAP mRNA stability is not a contributing factor to enhanced productivity
To further confirm that the elevated SEAP productivity was due to enhanced transcription
and not increased mRNA stability, we measured the stability of the SEAP mRNA
transcript in both control and miR-23 sponge clones after Actinomycin-D (Act-D)
treatment. Actinomycin-D treatment is used as a transcriptional block due to its high
affinity binding for DNA GC-rich duplexes at transcription initiation complexes and
preventing elongation by RNA polymerase (Kang, Park 2009). By inhibiting
transcription, the rate of mRNA decay, i.e. stability, can be determined. The two best
performing miR-23 sponge clones (4 and 6) and the best control clones (2 and 11) were
selected for stability assessment. RNA was harvested 6, 18 and 24 hs after the addition of
Act-D and assessed for SEAP mRNA levels using SYBR Green qPCR with β-actin as the
endogenous control. The cyclic threshold (Ct) values for each time point of each sample
is graphed below as a means to best depict the trend in mRNA stability across all
conditions (Fig. 3.68).
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
miR-NC spg 2
miR-NC spg 11
miR-NC spg 20
miR-23b spg 4
miR-23b spg 6
RQ
miR-23b expression
miR-23b
***
***
234
Figure 3.68: miR-23 sponge SEAP mRNA stability assay
Figure 3.68: qRT-PCR was used to evaluate the stability of the SEAP mRNA transcript in miR-23
sponge clones versus miR-NC sponge clones upon the addition of Actinomycin-D. Upon
transcriptional inhibition RNA was sampled 6, 18 and 24 hs later with the respective Ct values of
each clone graphed and slopes calculated as a means to assess the rate of decay.
We chose to represent mRNA stability as a function of the Ct from each time point and
calculate the slope of the line as a means to measure the change over time because
expressing changes in mRNA levels quantitatively was hindered by the fact that the
endogenous control (β-actin) displayed an accelerated decay rate to that of the SEAP
transcript. It can be observed that both miR-NC sponge clones and miR-23 sponge clones
behaved similarly reflecting both the similar SEAP production capacity previously
observed and the elevated SEAP transcript exhibited by both miR-23 sponge clones
indicated by the lower Ct values recorded for each time point. Furthermore, the slope
calculated for each clone miR-NC sponge 2 and 11 and miR-23 sponge 4 and 6 was
0.9493, 0.9743, 0.9873 and 0.9828, respectively. This observation indicated that the rate
of decay of the SEAP transcript was comparable between both the control and the test
sponge clones and that mRNA stability was not playing a role in the enhanced SEAP
activity seen in our CHO-SEAP clones engineered to stably deplete miR-23.
18
19
20
21
22
23
24
0 5 10 15 20 25 30
Ct
Time (Hrs)
Act-D treated
miR-NC 2
miR-NC 11
miR-23b 4
miR-23b 6
18
19
20
21
22
23
24
0 5 10 15 20 25 30
Ct
Time (Hrs)
Act-D treated
235
3.4.4 Cell size is not a contributing factor to clonal differences in productivity
Cell size has been a variable considered to play a part in the specific production capacity
of rCHO cells. Larger cells have been shown to possess inherently longer duplication
times (Kang et al. 2013a). This increased duplication time could correlate with a longer
period of G1 during the cell cycle or in the case of larger cells without increased genomic
size, an increased availability of biogenesis machinery.
Using the Cedex XS analyzer (see section 2.2.3.3 of Materials and Methods), the size
of the CHO-SEAP sponge clones were evaluated. miR-23 sponge clone 4 and 6 and miR-
NC sponge clone 2 and 11 were selected as these clones exhibited a similar SEAP
production profile. A size range of 11.3 µm to 12 µm was observed with miR-NC sponge
clone 11 falling at the higher end of this scale (Fig. 3.69). This indicated that clone cell
size was not contributing to the increase in productivity in miR-23 depleted clones.
Figure 3.69: Comparison of miR-23 sponge and miR-NC sponge clonal cell size
Figure 3.69: Using the Cedex XS analyzer, a Trypan blue-based cell imager, miR-23 sponge and miR-
NC sponge clones were assessed for their cell size. Cell size for each clonal population is represented
as an average value across the entire population.
miR-NC spg 2 miR-NC spg 11
miR-23b spg 6miR-23b spg 4
Ave = 12.095 µmAve = 11.48 µm
Ave = 11.3 µm Ave = 11.825 µm
236
3.4.5 miR-23 depleted CHO clones exhibited elevated Glutamate levels
Both miR-23a and miR-23b have previously been shown to target the mRNA of
glutaminase (GLS), an enzyme that mediates the conversion of glutamine to glutamate
(Gao et al. 2009), which has been proposed to play a role in the Warburg effect through
anaplerosis, the replenishment of TCA cycle intermediates through glutaminolysis (Dang
2010) potentially boosting oxidative metabolism, a process previously associated with
productivity (Templeton et al. 2013). For this reason, we measured the levels of secreted
glutamate within the supernatants of day 4 cultures (section 2.6.9 of Materials and
Methods).
Figure 3.70: Glutamate levels across sponge clones
Figure 3.70: Glutamate levels were measured from the supernatants of rCHO-SEAP clones
engineered with a miR-23 specific sponge (clone 4 and 6) and a non-specific control sponge (clone 2
and 11). Glutamate levels were calculated based on the average across both test and control groups
with statistical significance being calculated using a standard student t-test (p ≤ 0.05*).
The average glutamate levels across both clonal groups was observed to be increased by
65% (p ≤ 0.05, Fig. 3.70). This potentially suggested that glutamate levels were elevated
through de-repression of translation of the glutaminase enzyme. However, semi-
quantitative and quantitative PCR analysis did not indicate any change in the mRNA
0
0.005
0.01
0.015
0.02
0.025
0.03
0.035
NC-spg 23b-spg
g/L
*
237
levels of glutaminase (GLS), (Fig. 3.77 and Fig. 3.78). This however does not discount
miR-23 interfering with the translation of GLS. The conversion of glutamine to glutamate
by GLS generating ammonia (NH4) is not the only source of glutamate. Glutamate
dehydrogenase (GDH) can catalyse the reversible conversion of glutamate to α-
ketoglutarate (Spanaki, Plaitakis 2012). Glutamate oxaloacetate transaminase also
mediates the reversible conversion of oxaloacetate and glutamate to aspartate and α-
ketoglutarate (Perez-Mato et al. 2014). This would suggest that the levels of glutamate
can either contribute to cellular bioenergetics or balance the availability of TCA cycle
intermediates. It would be expected that the primary source of glutamate generation
would be intracellular, however, secreted glutamate levels would be expected to reflect
the internal glutamate levels.
3.4.6 miR-23 depletion impacts positively on CHO cell mitochondrial activity
Although the observation of increased availability of glutamate could not be definitively
attributed to enhanced GLS activity without the assessment of its mature protein levels,
we decided to assess the impact on mitochondrial function in a miR-23 depleted clone by
using high resolution respirometry (HRR) (see section 2.5.2 of Materials and Methods).
As glutamate provides essential electron donors (NADH and FADH2) critical to
mitochondrial function, we speculated that a miR-23 sponge clone would demonstrate an
increased mitochondrial activity, be it either directly or indirectly through miR-23’s
influence on GLS. Oxidative phosphorylation (OXPHOS) is a measure of oxygen
consumption culminating in the synthesis of ATP and the respiratory by-product, water.
By measuring oxygen consumption using HRR, we were able to compare the
mitochondrial function of intact cells under normal environmental conditions.
Additionally, permeabilization assays allowed for the interrogation of individual complex
components of the electron transport system (ETS, Fig. 3.71). A series of mitochondrial
substrates, uncouplers and inhibitors are listed in table 3.4.
238
Table 3.4: SUIT protocol substrates, uncouplers and inhibitors
Substrate Abbreviation Site of action Volume
(µL)
Final
Conc in
2mL
Pyruvate P Complex I (CI) 5 5 mM
Malate M CI 2.5 0.5 mM
Glutamate G CI 10 10 mM
Succinate S CI 20 10 mM
Ascorbate As CIV 5 2 mM
TMPD Tm CIV 5 0.5 mM
Cytochrome c C CIV 5 10 µM
ADP D CV 4 1 mM
Uncoupler
FCCP
F
Inner
Mitochondrial
membrane
1 µL steps
0.05 µM
steps
Inhibitors
Rotenone Rot CI 1 0.5 µM
Malonic Acid Mna CII 5 5 mM
Antimycin A Ama CIII 1 2.5 µM
Sodium Azide Az CIV 50 100 mM
Oligomycin Omy CV 1 2 µg/mL
239
Figure 3.71: Schematic diagram of the electron transport system (ETS) and its
component complexes
Figure 3.71: A schematic diagram showing the components of the electron transport system (complex
I, II, III and IV) culminating at the generation of ATP through Complex V. These complexes are
localised to the inner mitochondrial membrane and are responsible for transporting electrons from
donors (NADH and FADH2) along the chain ending in the generation of water form molecular oxygen
at complex IV. Additionally, generation of the proton gradient is also contributed to by these
complexes mediating the transfer of protons from the matrix to the inner membrane space. (Image
sourced from http://www.robertlfurler.com/category/systems-biology/).
3.4.6.1 Mitochondrial function of intact rCHO cells
Routine respiration (R) is a measure of mitochondrial oxygen consumption and normal
cellular homeostasis that is supported by exogenous substrates from the culture media.
Dissection of mitochondrial activity via the various components that make up the
mitochondrial electron transport system (ETS) cannot be measured in intact cells as the
plasma membrane is impermeable to various mitochondrial substrates and ADP.
Once, R has stabilised, ATP synthesis is inhibited through the addition of oligomycin
(ATP synthase-Complex V inhibitor). LEAK respiration (L) is a measure of cellular
240
substrate utilisation accommodated by the proton leak (intrinsic uncoupling) at maximum
mitochondrial potential. As ATP synthase has been inhibited, the electrochemical proton
gradient is not regulated as proton import into the mitochondrial matrix cannot occur. The
mitochondrial respiratory control (proton gradient exerting a regulatory electrochemical
backpressure) system is fully released or uncoupled by the addition of FCCP
(Carbonilcyanide p-triflouromethoxyphenylhydrazone) which cycles across the inner
mitochondrial membrane with the transport of protons and dissipates the proton gradient.
This allows the non-coupled state of ETS capacity to be determined irrespective of the
contribution from the proton pumps (CI, CIII and CIV).
Two clonal pairs were compared for their mitochondrial activity as intact cells (miR-NC
spg 2 and miR-23 spg 4: miR-NC spg 11 and miR-23 spg 6). Cells were cultured at an
initial density of 2 x 105 cells/mL in a 5 mL volume and grown under batch process
conditions for 72 h. Cells were counted and resuspended in a 4 mL volume of fresh culture
media at a final density of 1 x 106 cells/mL. Prior to inoculation of the OROBOROS
chambers, the oxygen sensors were calibrated using culture media. Given the limited
nature of the OROBOROS system, only two samples could be assessed at any one time,
therefore a test and corresponding control clone were assessed with results and statistics
calculated based on the values obtained from three replicates.
Routine (R) and LEAK (L) respiration of miR-23 depleted CHO cells was found to be
increased (p ≤ 0.05) as a function of the maximum ETS (E) determined by the addition
of FCCP (Fig. 3.72 B). Although the maximum ETS was observed to be lower in the case
of CHO cells exhibiting an increased volumetric yield (Fig. 3.72 A), respiratory control
ratios (RCRs) revealed that miR-23 depleted CHO cells produced a higher percentage of
their maximum energy capability as opposed to NC-sponge cells. This indicated that miR-
23 depleted clones under normal growth conditions have a more efficient mitochondrial
function. miR-23 sponge clone 4 was observed to perform at 54% of its maximum
capacity routinely (R) compared to 44% in miR-NC sponge clone 2 while leak was
observed to be 32% of its maximum compared to 27 (Fig. 3.72 B).
241
Figure 3.72: Mitochondrial function of intact miR-23 depleted rCHO cells clone 4
and miR-NC sponge 2
Figure 3.72: Comparison of mitochondrial activity based on the rate of oxygen consumption between
the miR-23 sponge clone 4 and the control sponge clone, miR-NC sponge 2. A) Represents the raw
values for the rate of oxygen consumption between both miR-sponge clones. B) Flux control ratios
(FCRs) are normalised to a common reference state after normalisation to residual oxygen (ROX).
R/E ratio is an expression of how close ROUTINE respiration operates to ETS capacity while L/E
follows the same trend. All calculations were made based on the individual values obtained over the
course of three experimental replicates (n = 3) with statistical significance calculated using a type 1
paired standard student t-test (p ≤ 0.05* and ≤ 0.01**).
Interestingly, when a second set of clones were evaluated for their mitochondrial activity,
miR-23 sponge clone 6 and miR-NC sponge clone 11, it was observed that there was a
significantly enhanced routine respiration by 22% (p ≤ 0.01) and maximum ETS by 13%
(p ≤ 0.05). Although there was an elevated ETS observed in this case, it did not translate
into a significant increase in the ratio of energy utilisation when normalised to the
maximum ETS (Fig. 3.73).
-10
0
10
20
30
40
50
60
70
80
90
Routine Leak ETS Rox
Ox
yg
en c
on
sum
pti
on
(p
mo
l/1
06
cel
ls
Raw Data
miR-NC 2
miR-23b 4**
0.00
0.10
0.20
0.30
0.40
0.50
0.60
0.70
R/E L/E
% o
f E
TS
mt-activity Rel. to ETS
miR-NC 2
miR-23b 4
*
*
-10
0
10
20
30
40
50
60
70
80
90
Routine Leak ETS Rox
Ox
yg
en c
on
sum
pti
on
(p
mo
l/1
06
cel
ls
Raw Data
miR-NC 2
miR-23b 4**
A
B
242
Figure 3.73: Mitochondrial function of intact miR-23 depleted rCHO cells clone 6
and miR-NC sponge 11
Figure 3.73: Comparison of mitochondrial activity based on the rate of oxygen consumption between
the miR-23 sponge clone 6 and the control sponge clone, miR-NC sponge 11. A) Represents the raw
values for the rate of oxygen consumption between both miR-sponge clones. B) Flux control ratios
(FCRs) are normalised to a common reference state. R/E ratio is an expression of how close
ROUTINE respiration operates to ETS capacity while L/E follows the same trend. All calculations
were made based on the individual values obtained over the course of three experimental replicates
(n = 3) with statistical significance calculated using a type 1 paired standard student t-test (p ≤ 0.05*
and ≤ 0.01**).
Both miR-23 depleted clones appeared to demonstrate an enhanced mitochondrial activity
be it when directly compared to the rate of oxygen consumption in the case of miR-23
spg clone 6, or be it through the more opportunistic utilisation of energy when compared
to their maximum ETS capacity, in the case of miR-23 sponge clone 4.
0.00
0.05
0.10
0.15
0.20
0.25
0.30
0.35
0.40
0.45
0.50
R/E L/E
% o
f E
TS
mt-activity Rel. to ETS
miR-NC 11
miR-23b 6
0
10
20
30
40
50
60
70
80
90
Routine Leak ETS Rox
Ox
yg
en
co
nsu
mp
tio
n (
pm
ol/
10
6cell
s)
Raw Data
miR-NC 11
miR-23b 6**
*
0
10
20
30
40
50
60
70
80
90
Routine Leak ETS Rox
Ox
yg
en
co
nsu
mp
tio
n (
pm
ol/
10
6cell
s)
Raw Data
miR-NC 11
miR-23b 6**
*
243
Ideally, assessing and comparing the various combinations of test versus control sponge
clones would have been both desirable and beneficial, however, the limited number of
chambers in the OROBOROS system would make this a time consuming endeavour
which unfortunately could not be explored.
3.4.6.2 Mitochondrial function of permeabilised rCHO cells
Next we sought to evaluate if this enhanced mitochondrial function was as a result of the
hyperactivity of individual complex components that make up the electron transport
system (ETS). Permeabilised experimental evaluation of mitochondrial function was
carried out in mitochondrial respiration media (MiR05) which contains no exogenous
substrates. As a result, routine mitochondrial respiration cannot be assessed within
permeabilised mitochondrial assays as this is a natural state of respiration dependent on
exogenous sources.
The first exercise was to determine the concentration of Digitonin (Dgt) to achieve full
membrane permeabilization. Each chamber was inoculated with 2 mL of MiR05 buffer
containing 1 x 106 cells/mL. Cells were initially treated with 1 µL Rotenone (Complex I
(C1) inhibitor), 20 µL Succinate (CII substrate) and 4 µL ADP (precursor for ATP
production – see table 3.4 for final concentrations). In intact cells, CI inhibition via
rotenone addition suspends respiration through CII because succinate cannot cross the
cell membrane and succinate production through the TCA cycle is stopped when NADH
cannot be oxidised. Therefore, inhibition of CI in this instance allows us to assess cell
membrane permeability with mitochondrial respiration via CII being dependent purely on
exogenous sources of succinate. After inoculation of both chambers, Digitonin was
titrated 1 µL at a time with one chamber remaining as a control. As the cell membrane
becomes permeabilised, succinate moves into the mitochondria thus stimulating oxygen
consumption through CII activity. After each addition of Dgt, the rate of oxygen
consumption is allowed to stabilise before the addition of the next µL. We determined
that our rCHO cells reached a plateaux of permeability upon the addition of 3-4 µL of
Dgt (Fig. 3.74). Subsequent permeabilization experiments were carried out by initially
adding 4 µL of digitonin.
244
Figure 3.74: Cell membrane permeabilization optimisation using Digitonin
Figure 3.74: The addition of Digitonin (Dgt) was titrated in order to assess the optimal concentration
that would fully permeabilise the CHO cell membrane. As succinate cannot cross the cell membrane,
intact cells are initially dosed with the complex II substrate, succinate (S). Dgt is subsequently titrated
until maximum oxygen consumption has been achieved through full permeabilization of the cell
membrane allowing succinate and ADP to fully cross into the mitochondria.
The systematic evaluation of the activity of individual mitochondrial membrane
components that contribute to cellular respiration was assessed when miR-23 sponge
clone 4 and miR-NC sponge clone 2 were compared (Fig. 3.75 A and B). Part A of Figure
3.75 shows the real-time change in oxygen consumption of both control and test clones
with each substrate/inhibitor/uncoupler being indicated whereas part B demonstrates the
average values and statistics across a triplicate run.
A 29% increase in mitochondrial respiration was observed upon addition of
glutamate/malate as substrates for Complex I. However a significant 26% increase (p ≤
0.05) in OXPHOS in miR-23 depleted CHO cells was observed when ADP was no longer
the limiting factor (Fig. 3.75 BADP).
The addition of cytochrome c confirmed that in both cases mitochondrial membrane
integrity was not compromised either through the permeabilization process itself or as a
result of mitochondrial hyperactivity, in the case of miR-23 sponge cells.
0
10
20
30
40
50
60
70
80
90
100
0 1 2 3 4 5
Ox
yg
en
Co
nsu
mp
tio
n [
pm
ol/
10
6cell
s
Digitonin (µl)
No Dgt
Dgt
245
OXPHOS maintained the increased of ~ 29% (p ≤ 0.05) upon the addition of succinate, a
substrate for Complex II, suggesting that complex I was responsible for the observed
increase in mitochondrial respiration. However, upon acquisition of the ETS, complex I
was inhibited by the addition of Rotenone (Rot) resulting in a marginal drop in oxygen
consumption when compared to the addition of succinate (CI and CII). This suggested
that both complex I and complex II are mediating a ~30% (p ≤ 0.05) increase in cellular
respiration in miR-23 depleted CHO cells but do not appear to display an additive
advantage. However, complex II on its own in an uncoupled state maintained a 44%
higher (p ≤ 0.05) activity upon complex I inhibition (Fig. 3.75 BRot).
Subsequent inhibition of complex II and III followed by the activation of CIV through
the addition of the electron donor TMPD resulted in a marginal drop in oxygen
consumption in miR-23 depleted cells (Fig. 3.75 BTMPD).
Dissection of the mitochondrial membrane complexes suggested that complex I and II
contribute to a higher degree of respiration at levels comparable to each other but do not
display synergy when co-stimulated possibly due to a saturation point centred on the
downstream electron chaperone complexes III and IV.
246
Figure 3.75: Mitochondrial permeabilization assay on miR-23 sponge clone 4 versus
miR-NC sponge clone 2
Figure 3.75: Mitochondrial activity of permeabilised rCHO cells (miR-23 sponge clone 4 and miR-
NC sponge clone 2) using a SUIT protocol. A) Is a representation of a real-time read out of oxygen
consumption from one of the replicates (n = 3) indicating the addition of each
substrate/inhibitor/uncoupler over the time course (h: min). Oxygen consumption is allowed to
stabilise before the addition of Dgt. At each step, oxygen consumption is allowed to stabilise before
the addition of the next substrate/inhibitor/uncoupler. Glutamate/Malate (GM) is added to stimulate
CI activity while ADP (D) further stimulates OXPHOS through CI by relieving ADP availability as
a limiting factor. Cytochrome c (C) assessed the integrity of the mitochondrial membrane. Succinate
(S) additionally stimulates the activity of CII while the uncoupler FCCP (F) assessed the ETS though
titration at 2 µl intervals. Rotenone (Rot) addition inhibits CI allowing OXPHOS through CII and
CIII. Malonic acid/Antimycin A (Mna/Ama) inhibits CII and CIII, respectively. CIV is then
stimulated by the artificial electron donor TMPD in the presence of Ascorbate (AsTm) which serves
to maintain TMPD in a reduced state. Finally complex CIV is inhibited through the addition of
Sodium Azide (Az). B) Represents the same data normalised to ROX, with averages across all
replicates being shown. The states of mitochondrial activity is broken up in LEAK, OXPHOS and
ETS with statistical analysis calculated using a type 1 paired student t-test (p ≤ 0.05*).
LEAK OXPHOS ETS
CI CI + II CII CIV
Range [h:min]: 1:10
2:081:561:451:331:211:10
O2
Flo
w p
er
ce
lls
(A
) [p
mo
l/(s
*Mill)
]
160
128
96
64
32
0
Dgt GM D c S F1 F2 F3 F4 R MnaAma AsTm AzDgt GM D C S F1 F2 F3 F4Rot MnaAma AsTm Az
A
B
247
3.4.7 Mitochondrial content
Although mitochondrial function based on oxygen consumption was performed on a fixed
and equal number of cells (1 x 106 cells/mL) it does not account for the potential intrinsic
change in mitochondria per cell that could ultimately contribute to the elevated
mitochondrial activity. Several biomarkers have been assessed for their utility as
candidates whose activity correlate well with mitochondrial content including cardiolipin,
citrate synthase (CS) and complex I, in that order (Larsen et al. 2012). However, we
decided to use a fluorescent-based assay, MitoTracker Deep Red (MTDR, section 2.5.1
of Material and Methods), which is routinely used to evaluate mitochondrial content
(Kitami et al. 2012). Cells were cultured in a similar fashion to those that were evaluated
for mitochondrial activity using the OROBOROS in the previous section.
Sample fluorescence signals are affected by signal levels, auto-fluorescence, non-specific
staining, electronic noise, and optical background from other fluorochromes. These
factors can contribute to the noise and/or signals observed and increase the width of a
negative population (i.e. variance or standard deviation). Stain index (SI) normalization
was used to adjust for the influence of such factors. SI is defined as D/W, where D is the
difference between positive and negative populations and W is equal to 2 S.D. of the
negative population.
This data indicates that the enhanced mitochondrial function observed in the case of miR-
23 sponge clones is not as a result of increased mitochondrial mass/content (Fig. 3.76).
FACS-based assessment of mitochondrial content using MitotrackerTM Deep red does not
discriminate between mitochondrial mass and content. Immuno-histochemical staining
would provide further information as to the nature of mass being attributed to content
variation. Furthermore, the emission of MitotrackerTM DR at 665 nm did not interfere
with the inherent emission signal at 506 nm for the GFP positive sponge clones.
248
Figure 3.76: Mitochondrial mass/content assessment using FACS analysis
Figure 3.76: Assessment of mitochondrial content in stable miRNA sponge expressing CHO cells.
Mitochondrial content of control miR-NC2, -NC11 sponge and miR-23-6 and miR-24-4 expressing
CHO cells were examined by MTDR staining and flow cytometry. Data were obtained from ~10,000
events and stain index used to normalize fluorescent signals for each sample. A representative
histogram shows MitotrackerTM deep red signals for stained and unstained test and control samples.
3.4.8 Assessment of miR-23 target transcript levels
Several candidate genes were chosen for evaluation of their endogenous transcript levels
for two reasons. Firstly, the enzyme glutaminase (GLS) has been validated to be a target
of miR-23a/b which in the case of our stable miR-23 sponge CHO cells could potentially
contribute to the elevated glutamate levels (Gao et al. 2009). Furthermore, another
interesting target of miR-23 is the mitochondrial transcription factor A (TFAM) which
has been shown to play a critical role in the transcriptional initiation of mitochondrial
encoded genes (Jiang et al. 2013b). Secondly, increased mitochondrial activity has been
correlated with enhanced global transcriptional activity (das Neves et al. 2010). From this
perspective we selected this panel of genes including another validated target of miR-23,
XIAP (X-linked inhibitor of apoptosis), for the assessment of their mRNA levels (Siegel
et al. 2011). As miRNA function predominantly lies at the post-transcriptional level, the
observation of unchanged transcript levels of the aforementioned targets does not
discount the possibly of miR-23 depletion impact on the mature protein levels. However,
APC-A
Cou
nt
-101
102
103
104
105
0
35
70
105
140
ParentmiR-23 spg 2miR-23 spg 6miR-NC spg 2miR-NC spg 11
249
if enhanced mitochondrial activity was having an effect on global gene transcription, it
was suspected that all three genes could be elevated at an mRNA level.
Semi-quantitative PCR was carried out on all three selected targets of miR-23 using the
designed primers listed in the appendix (Table 2.2 of Materials and Methods) using
DreamTaq (Thermo Scientific) following the protocol in section 2.4.3 of Materials and
Methods.
Figure 3.77: Semi-quantitative PCR for miR-23 targets
Figure 3.77: Semi-quantitative assessment of the mRNA transcript levels for a select panel of
previously validated targets of miR-23 (GLS, TFAM and XIAP). A moderate number of cycles, 25,
was used so as not to saturate the amplification signal thus allowing visual identification of potential
fluorescent changes. Β-actin was used as an endogenous control to ensure efficient sample loading.
All PCR amplicons were pre-designed to be small, in the range of 150-200 bp, so were ultimately ran
on a high percentage (2%) agarose gel stained with ethidium bromide and ran slowly for the
acquisition of clean bands.
250
There appeared to be no apparent changes in the abundance levels of the mRNA
transcripts for XIAP and TFAM in all rCHO-SEAP clones from both non-specific and
miR-23 sponge panels that were selected for clonal screening (Fig. 3.77). There did
appear to be a reduced level of expression for the transcript of glutaminase (GLS).
However, minute changes in expression can be difficult to observe at best even if an
optimal number of cycles has been utilised.
We next sought to evaluate the abundance of transcript levels for the same targets using
quantitative real-time RT-PCR (see section 2.4.4 of Materials and Methods). Clones
chosen to take part in this screen were the two best producing miR-23 sponge clone 4 and
6 and the two best non-specific control clones, miR-NC sponge clone 2 and 11.
Little to no change was observed for the expression of GLS and TFAM when qPCR was
carried out (Figure 3.78). A moderate increase in the expression of XIAP was observed
in the case of miR-23 sponge clone 6 and for a single control miR-NC sponge clone 11
(Figure 3.78XIAP), potentially a result of clonal variation. This observation does not
discount the possibility that miR-23 depletion is having an effect on one or all of these
validated targets due to the nature of miRNA mode of action, translational inhibition.
To fully understand the potential influence of miR-23 depletion on these target genes,
analysis on the protein level would have to be carried out through western blot. Label-
free mass-spectrometry will be explored in these stable rCHO-SEAP clones in the next
section as a means to identify potential targets of miR-23 and to isolate pathways
potentially contributing to the observed phenotype of enhanced target protein productivity
with elevated mitochondrial respiration. For this reason, identification of protein
abundance of the three targets was not explored at this time through western blot analysis.
251
Figure 3.78: qPCR of miR-23 validated targets GLS, TFAM and XIAP
Figure 3.78: Quantitative evaluation of changes in the expression of the mature transcript of a panel
of validated targets of miR-23, GLS, TFAM and XIAP. Fast SYBR® Green was used to evaluate the
expression of three targets previously assessed at a semi-quantitative level. Data was normalised to
one of the two control clones (miR-NC spg 2). Relative quantification was calculated using the 2–ΔΔCt
method.
0
0.2
0.4
0.6
0.8
1
1.2
1.4
NC-spg 2 NC-spg 11 23b-spg 4 23b-spg 6
RQ
Glutaminase
GLS
0
0.2
0.4
0.6
0.8
1
1.2
1.4
NC-spg 2 NC-spg 11 23b-spg 4 23b-spg 6
RQ
TFAM
TFAM
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
2
NC-spg 2 NC-spg 11 23b-spg 4 23b-spg 6
RQ
XIAP
XIAP
252
It has been demonstrated so far that stable depletion of miR-23 improves the productivity
of SEAP by an average of 3-fold when compared to non-specific controls. Further
evaluation found that SEAP productivity may be attributed in part to an increase in the
level of SEAP mRNA. Furthermore, mRNA stability and cell size were ruled out as
contributing to this.
miR-23 has been strongly implicated in mitochondrial function both through the
reprogramming of the “Warburg effect” through anapleurotic replenishment of TCA
cycle intermediates such as α-ketoglutarate (αKG) via enhanced translation of the enzyme
glutaminase (GLS) (Liu et al. 2012) and additionally through the post-transcriptional
attenuation of the mitochondrial transcription factor A (TFAM) (Campbell, Kolesar &
Kaufman 2012). Although the mRNA levels of these genes remained unchanged in our
clones, it does not rule out the possibility of their mature protein products being
influenced especially when considering the elevated glutamate levels observed in the
miR-23 depleted clones. Using high resolution respirometry (HRR), oxygen consumption
was assessed as a measure of mitochondrial activity. A ~30% increase in oxygen
consumption was observed in the miR-23 sponge cells attributed to Complex I and
Complex II stimulation. Both complexes contributed to this enhanced mitochondrial
function independently of each other but did not appear to act synergistically, potentially
suggesting a bottleneck downstream. Using a mitochondria-specific fluorescent dye, we
managed to rule out an increase in cellular mitochondrial mass/content as a contributing
factor to this enhanced activity.
The next section explores the discovery of potential targets of miR-23 that could have a
role in mediating this observed influence on CHO cell bioenergetics in addition to
identifying molecular pathways directly or indirectly influenced by miR-23 depletion.
253
Section 3.5
miR-23 target identification and
pathway analysis
254
3.5 microRNA target identification
Previous characterisation has so far identified that stable miR-23 depletion results in
increased specific SEAP productivity without any impact on cellular growth thus
ultimately enhancing volumetric SEAP yield. Across a panel of clones the enhanced
SEAP activity appears to be somewhat attributable to the increase in transcription of the
SEAP mRNA. Previous characterisation from the literature has implicated miR-23 to be
involved in cellular bioenergetics through the regulation of targets associated with
mitochondrial function such as GLS and TFAM. However, analysis of the expression of
these transcripts using qPCR revealed no difference in the case of miR-23 depletion in
CHO cells. Given the nature of miRNA, this observation does not rule out the possibility
that the mechanism of action of miR-23 is being mediated through blocking translation.
In order to attempt to identify how miR-23 depletion selectively appears to enhance SEAP
transcription and influence mitochondrial functional, we set out to identify potential
targets of miR-23 through a novel process of biotin-labelled microRNA mimics and
Label-free Mass Spectrometry on rCHO-SEAP clones stably engineered to deplete miR-
23. It was hoped that the global mRNA targeting capacity would be identified in the case
of biotin-tagged mRNA pull down thus indicating potential pathways under miR-23
regulatory control as well as high priority candidate targets, while proteomic profiling
was intended to unveil a snap shot of mature molecular pathways under functional
interference from miR-23, be it either directly or indirectly.
3.5.1 Identification of global mRNA targets through Biotinylated miRNA tags
A recently described method of identifying miRNA targets is the use of biotinylated-
labelled miRNA mimics (Orom, Lund 2007) and successfully reported for miR-34a (Lal
et al. 2011). This technique exploits the use of miRNA mimics (pre-miRs) whose 3’-end
has been covalently modified to carry a biotinylated tag (Fig. 3.79 A). Streptavidin,
isolated form Streptomyces avidinii and immobilised to sepharose beads has a high
binding affinity for biotin or biotinylated components (Fig. 3.79 A).
255
Figure 3.79: Schematic diagram of biotin-labelled miRNA-mediated mRNA pull
down
Figure 3.79: A) Shows a mature miRNA with a chemically modified 3’- end (red arrow) to include a
linked Biotinylated group as well as the streptavidin-agarose beads with high affinity binding sites
for biotin B) Represents the general flow diagram for miRNA target enrichment. RNA is isolated
from streptavidin-agarose beads through the Tri Reagent protocol C) During the incubation period
of cell lysate with streptavidin-agarose beads mRNAs bound to the biotin labelled miRNA will remain
bound to the SA beads and enriched for after clearing of the cell lysate.
By incubating CHO cell lysates that have been lysed though the utilisation of a gentle
lysis buffer so as not to disturb miRNA-RISC bound target mRNAs in the presence of
streptavidin beads, it is possible to wash away the majority of non-bound mRNAs and
enrich for those that have a high binding affinity for the mature biotin-tagged miRNA.
Using microarray technology (section 2.4.10 of Material and Methods), it is possible to
identify enrichment of a panel of mRNA targets whose identification is potentially a
function of their association with the pre-tagged miRNA of interest.
A
B
5’-
C
Streptavidin-agarose beadsMature miRNA with 3’- biotin tag
Cells transfected with
biotin-labelled miRNA
Gentle Lysis
buffer
Incubate with
Streptavidin-agarose
beads
miRNA target
enrichment
Centrifuge and clear
supernatant
256
Both miR-23b and its counterpart passenger strand miR-23b* were selected for mRNA
target identification through this novel method in an attempt to determine firstly what
targets and molecular networks are under their influence and secondly if there is
complementary, antagonistic or unique cellular process under their individual regulation.
3.5.1.1 Evaluation of biotin-miRNA workflow
Prior to committing to a full scale microarray analysis, validation of the protocol was
assessed using a validated target as previously reported for the miRNA bantam (Hid gene)
(Orom, Lund 2007) and miR-34a (CDK4/6) (Lal et al. 2011).
We undertook some validation steps using qPCR to assess the utility of this method using
validated miR-23b* targets such as proline dehydrogenase (PRODH) (Liu et al. 2010a).
However, a challenge presented was the lack of validated targets of miR-23b* other than
PRODH. A panel of potential targets of miR-23b* was selected by generating a list of
mRNAs that were anti-correlated to the expression of miR-23b* from the datasets
acquired during the DE of miRNA in fast and slow growing CHO cells (section 3.1 of
results and (Clarke et al. 2012)), unveiling 319 mRNAs.
Next, the miRWalk prediction algorithm was utilised to generate two lists of potential
miR-23b* targets, one predicted and one validated (Dweep et al. 2011a). From here, both
lists were analysed for any overlap with the list of mRNAs found to be anti-correlated
with miR-23b* expression. 11 mRNAs were found to overlap with the list of 856
predicted targets while none were found to overlap with the list of 98 validated targets
(Table 3.5). The search criteria for “Validated targets” uses an algorithm that searches
abstract and paper content for commands such as “miR-23b*” and associated gene names.
Therefore, this list of 98 validated targets would not necessarily be accurate and reading
the paper content would be essential to establish if they could truly be considered
“validated”.
257
Table 3.5: miRWalk Predicted targets of miR-23b* overlapping with targets anti-
correlated to miR-23b* expression
Although PRODH did appear in the list of validated targets, it was not predicted to be a
target of miR-23b* nor did it appear to be anti-correlated with miR-23b* from the
miRNA:mRNA profile.
Biotin-labelled miR-23b* and miR-NC were transfected into CHO-SEAP cells at the
usual concentration of 50 nM in 1 x 105 cells/mL (see section 2.3.2.2 of Materials and
Methods). After 72 h, cells were harvested and processed according to the protocol
described by Ørum and Lund (Orom, Lund 2007) and described in detail in section 2.7
of Material and Methods. Two miR-23b* biotin –tagged mimics were designed, miR-
23b*-1 and miR-23b*-2 (natural). miR-23b*-1 comprises a duplex miRNA with a
complementary strand designed to promote strand selection of miR-23b* (Appendix
Figure A8). miR-23b*-2 (natural) was specifically designed to mimic the wild-type
hairpin and would result in the selection of the mature miR-23b. When both biotin-tagged
miRNA were transfected separately, qRT-PCR was carried out for both miR-23b*-1 and
miR-23b*-2 (natural) (see section 2.4.8 of Materials and Methods) probing for miR-
23b and the miR-23b*.
When miR-23b* levels were assessed, an increase was observed of ~600-fold for miR-
23b*-1 and ~12,000-fold for miR-23b*-2 (natural) in bead enriched samples (Fig. 3.80
A). Furthermore, a ~2000-fold increase of miR-23b* was detected for miR-23b*-1 and
~9,000-fold for miR-23b*-2 (modified) in the input samples prior to bead exposure (Fig.
3.80 B). Input samples, are samples taken after gentle lysis but prior to incubation with
streptavidin beads.
Top 10 predicted targets Overlapping targets Selected Targets
DNAJC3 Rapgef1 Rapgef1
GRIP1 BSG CSTA
CRYBB2 Plod1 DNAJC3
CTSA C6orf89 GRIP-1
HCG9 GGA2 PRODH
PGK1 BEST1 PGK-1
SMPD1 ACYL BEST-1
STX4 YARS CALU
ACLY RARRES2 GGA2
SPRN CALU YARS
Pomt2
258
Figure 3.80: Quantification of miR-23 and miR-23b* by qPCR from streptavidin
pull-down
Figure 3.80: qRT-PCR data assessing the mature levels of miR-23b and miR-23b* when biotin-
labelled miR-23b*-1 and miR-23b*-2 (modified) were transfected into CHO-SEAP cells. miR-23b*
is probed for in A) streptavidin captured samples and B) Input samples taken after gentle lysis but
before incubation in streptavidin beads. miR-23b is probed for in C) streptavidin captured samples
and D) Input samples.
As an additional level of quality control, miR-23b levels were assessed in the same
samples. No fold change of miR-23b was observed in either streptavidin captured or input
samples in the case of miR-23b*-1 as expected as this biotin-labelled pre-miRNA is
designed to exclusively express miR-23b* as its complementary strand does not contain
mis-matches therefore has a different sequence to miR-23b (Fig. 3.80 A and B).
Furthermore, when miR-23b levels were measured in the case of the miR-23b*-2
(modified) transfectant, a ~14,000-fold and 1,000-fold increase was observed for
streptavidin captured and input samples, respectively (Fig. 3.80 C and D). This
enrichment for miR-23b in the case of the tailored miR-23b*-2 (natural) biotin-tagged
duplex was expected as it should mimic a naturally occurring miR-duplex. The miR-
0
2000
4000
6000
8000
10000
12000
14000
miR-23b*-1 miR-23b*-2 NC-Biotin Cells
Re
lati
ve Q
uan
tita
tio
n (
RQ
)
miR-23b* (Bead Pulldown)
miR-23b*
0
1000
2000
3000
4000
5000
6000
7000
8000
9000
10000
miR-23b*-1 miR-23b*-2 NC-Biotin Cells
Re
lati
ve Q
uan
tita
tio
n (
RQ
)
miR-23b* (Input)
miR-23b* (Input)
0
2000
4000
6000
8000
10000
12000
14000
16000
miR-23b*-1 miR-23b*-2 NC-Biotin Cells
Re
lati
ve Q
uan
tita
tio
n (
RQ
)
miR-23 (Bead Pulldown)
miR-23 (Beads)
0
200
400
600
800
1000
1200
miR-23b*-1 miR-23b*-2 NC-Biotin Cells
Re
lati
ve Q
uan
tita
tio
n (
RQ
)
miR-23 (Input)
miR-23 (Input)
A B
C D
259
23b*-1 biotin-tagged pre-miR was ultimately selected for experimental evaluation as it
best represents the processing steps that culminate at RISC loading.
Next, we assessed if successful pull-down of the list of potential miR-23b* targtes could
be observed in samples transfected with miR-23b*-1 and -2 and exposed to streptavidin
beads. The amount of RNA pull-down was approaching the sub-nanomolar concentration
when assessed using a NanoDrop, however, enough template was retrieved that enabled
reverse transcription and subsequent PCR analysis. Of the list of 10 candidates selected
for evaluation, Pgk-1 resulted in the most consistent enrichment. miR-23b*-1 proved to
enrich for Pgk-1 compared to miR-23b*-2 in cell lysates exposed to streptavidin (Fig.
3.81 Beads). An inconsistency can be observed across input samples possibly due to the
quality of the RNA harvested.
Figure 3.81: Enrichment for Pgk-1 in miR-23b*-biotin pull-down
Figure 3.81: Analysis of Pgk-1 expression in CHO-SEAP cells transfected with biotin-tagged pre-
miRs, miR-23b*-1 and miR-23b*-2, and incubated with streptavidin beads (Beads). RNA samples
were taken prior to exposure to streptavidin (Input).
β-actin
Pgk-1
Beads
Input
260
Assessment of Pgk-1 pull-down by miR-23b*-1 biotin-tagged pre-miR upon several
repeats demonstrated inconsistent yet strong indications that Pgk-1 specific binding and
enrichment was occurring (Fig. 3.82).
Figure 3.82: Enrichment for Pgk-1 in miR-23b*-1 biotin-labelled transfected CHO-
SEAP cells
Figure 3.82: pPCR analysis to measure enrichment of the potential target of miR-23b*, pgk-1 (white
arrow), in CHO-SEAP cells transfected with miR-23b*-1. “Pull-down” indicates samples incubated
in streptavidin beads while “input” indicates RNA sampled prior to bead incubation.
3.5.1.2 Microarray identification of mRNA targets specific for miR-23b and miR-
23b* using biotin-labelled mimics
Following initial optimisation, biotin-labelled miR-23b and miR-23b*-1 pre-miRs were
transfected into parental CHO-SEAP cells and processed according to section 2.7.5 of
Materials and Methods. As 12.5 µg of aRNA (antisense RNA) was required for loading
of CHO-specific microarray chips, an amplification process (MessageAmpTM II aRNA
amplification, section 2.4.9 of materials and methods) was undertaken prior to
processing of the aRNA (section 2.4.10 of Materials and Methods).
Subsequent to generating aRNA, samples were hybridized to CHO-specific microarrays.
Across all samples (cells only, miR-NC-biotin, miR-23b-biotin and miR-23b*-1-biotin),
100 bp
300 bp
Pulldown Input
Pgk-1
261
1933 annotated genes were identified as being present, however, none demonstrated a
significant fold change or enrichment between conditions. One of the first quality control
tests undertaken in microarray analysis is the use of hierarchal clustering. This is an
unbiased means by which a dataset is grouped into sample families based on a similar
gene expression pattern. This revealed that there were no similarities between samples of
the same group (Fig. 3.83).
Figure 3.83: Cluster Dendrogram of microarray results
Figure 3.83: Dendrogram demonstrating the unsupervised hierarchal clustering of samples based on
gene expression.
Any changes observed between biotin-tagged miRNAs compared to biotin-tagged non-
specific controls equated to inherent background noise that would ordinarily be observed
262
between replicate samples. Although this method of miRNA:mRNA identification has
proved valuable and successful in the past, in this case, no candidates were generated
from our experiment.
The reasons for this was not immediately apparent so we decided to check bead-elutes
for enrichment of both miR-23b and miR-23b* in case the capture step failed.
Figure 3.84: miR-23b/23b* enrichment in main experimental run
Figure 3.84: qRT-PCR analysis using singleplex PCR specific for miR-23b* (blue) and miR-23b (red)
was carried out on the streptavidin incubated biotin-labelled pre-miRNA for the mature miR-23b
and miR-23b*.
Singleplex qRT-PCR carried out on samples captured for microarray analysis revealed
that there was a ~14-fold enrichment of miR-23b* with no apparent enrichment of miR-
0
2
4
6
8
10
12
14
16
Cells Biotin NC Biotin miR-23b Biotin miR-23b* Biotin
RQ
Treatment
miR-23b* enrichment in biotin pull-down
miR-23b*
0
0.2
0.4
0.6
0.8
1
1.2
1.4
Cells Biotin NC Biotin miR-23b Biotin miR-23b* Biotin
RQ
Treatment
miR-23b enrichment in biotin pull-down
miR-23b
263
23b (Fig. 3.84). This latter result would explain the lack of clustering observed in figure
3.83 for miR-23b samples but, miR-23b* at least, seemed to have been successfully
pulled down although not to as high a magnitude as previously achieved.
3.5.2 Identification of global protein targets using Label-free Mass Spectrometry
As miRNAs ultimately influence translation, we sought to carry out label-free mass
spectrometry as a means to identify the potential finite changes in the proteome elicited
by the stable depletion of miR-23 through sponge intervention. In addition, we attempted
to identify potential molecular networks and/or biological processes associated with
mitochondrial function.
3.5.2.1 Differentially expressed proteins in miR-23 depleted clones
Differential protein expression analysis was performed on two miR-23 sponge clones
(clone 4 and 6) and two miR-NC sponge clones (clone 2 and 11), the same clones that
were evaluated for mitochondrial function in section 3.4.6. CHO cells were cultured under
normal batch conditions in an identical manner as had been previously described for
mitochondrial function assays and harvested for whole cell proteome analysis after 72 hs.
A maximum of 500 µg of protein was prepped using a Ready-prep 2D clean-up kit (Bio-
Rad, #163-2130, and see materials and methods section 2.8.3) as a means to further
concentrate, if necessary, and remove any contaminating salts, reagents and buffers from
the lysis process. 10 µg of protein was then subsequently digested with Lys-c and trypsin
overnight to cleave all protein into polypeptides for mass spec identification.
41 proteins were found to be differentially expressed (25 up-regulated and 16 down-
regulated) in miR-23 depleted CHO-SEAP clones when compared to their non-specific
control counterparts (Table 3.6). Proteins are listed in order of MASCOT score output
which overlaps the mass spectral data of the target peptides with that of the reference
database, including the number of identifications (numbers of peptides associated with a
single protein). MASCOT also incorporates a false discovery rate (FDR) which compares
the same mass spectral data of the target peptides to that of a randomised reference
database of similar length as a means to identify the probability of identifications by
264
chance. Statistical significance is then assigned to each identified protein after
normalisation (Anova value) to determine that the observed fold-changes demonstrate
low variation within replicates and/or pooled clones in the case of our experiment.
265
Table 3.6: DE proteins identified through label-free mass spectrometry in miR-23 (test) depleted CHO-SEAP clones
Up in Test
Control Test
Down in Test
gi|860908 6 271.74 7.74E-03 1.5 vimentin [Cricetulus griseus] [MASS=44524] VIM 2.91E+06 1.94E+06
0.67
gi|344235837 3 (2) 188 3.36E-05 1.74 Glutathione S-transferase Mu 6 [Cricetulus griseus] [MASS=86559] GSTM6 2.15E+05 3.75E+05
1.74
gi|344236438 3 148.52 6.85E-05 2.07 Bifunctional aminoacyl-tRNA synthetase [Cricetulus griseus] [MASS=154896] EPRS 4.09E+05 1.98E+05
0.48
gi|344257633 2 145.81 2.72E-08 1.98 Lactadherin [Cricetulus griseus] [MASS=16152] MFGE8 1.41E+05 2.79E+05
1.98
gi|344255736 3 144.97 3.12E-04 2.16 Non-muscle caldesmon [Cricetulus griseus] [MASS=61921] CALD1 2.09E+05 4.52E+05
2.16
gi|354487570 2 133.91 6.64E-05 1.85 PREDICTED: ATP-binding cassette sub-family F member 1-like [Cricetulus
griseus] [MASS=95446]
ABCF1 1.39E+05 7.53E+04
0.54
gi|344237556 2 (1) 132.94 0.02 1.52 Glyceraldehyde-3-phosphate dehydrogenase [Cricetulus griseus]
[MASS=36889]
GAPDH 4.80E+05 3.15E+05
0.66
gi|354479271 2 118.37 8.74E-07 2.49 PREDICTED: perilipin-4-like [Cricetulus griseus] [MASS=184216] PLIN4 1.06E+05 2.63E+05
2.48
gi|354507545 2 (1) 98.26 4.80E-03 1.56 PREDICTED: glutathione S-transferase Mu 1-like [Cricetulus griseus] GSTM1 1.32E+04 2.07E+04
1.57
gi|344238611 1 91.09 2.48E-03 1.68 Isocitrate dehydrogenase [NADP] cytoplasmic [Cricetulus griseus]
[MASS=12467]
IDH1 2.58E+04 4.33E+04
1.68
gi|344244293 1 89.46 4.33E-07 2.73 LETM1 and EF-hand domain-containing protein 1, mitochondrial [Cricetulus
griseus] [MASS=79070]
LETM1 2.71E+04 7.39E+04
2.73
Description Average
Normalised
Abundances
Accession Peptides Score Anova
(p)*
Fold Tags Gene I.D.
266
gi|354474417 1 86.11 5.14E-05 1.54 PREDICTED: ras GTPase-activating protein-binding protein 1 [Cricetulus
griseus] [MASS=51797]
G3BP1 1.07E+05 6.96E+04
0.65
gi|354473301 2 85.25 6.20E-04 1.54 PREDICTED: septin-9 [Cricetulus griseus] SEPT9 1.76E+05 2.71E+05
1.54
gi|344251753 1 85.03 0.02 317.52 Heat shock cognate 71 kDa protein [Cricetulus griseus] [MASS=22696] HSPA8 5.32E+05 1676.73
0.00
gi|4504445 1 78 0.04 2877.1 heterogeneous nuclear ribonucleoprotein A1 isoform a [Homo sapiens] HNRNPA1 2.24E+05 77.98
0.00
gi|344254200 1 77.24 1.48E-04 1.84 Prolyl 4-hydroxylase subunit alpha-1 [Cricetulus griseus] [MASS=58084] P4HA1 7.12E+04 1.31E+05
1.84
gi|354489184 1 72.31 1.05E-03 1.65 PREDICTED: lysosomal alpha-glucosidase-like [Cricetulus griseus] GAA 3.96E+04 6.55E+04
1.65
gi|344255594 1 71.25 4.95E-04 1.76 Calcium/calmodulin-dependent protein kinase type II delta chain [Cricetulus
griseus] [MASS=25006]
CaMK2D 8.60E+04 1.51E+05
1.76
gi|344248788 1 65.91 0.04 2.50E+04 Eukaryotic translation initiation factor 5A-1 [Cricetulus griseus]
[MASS=7244]
EIF5A1 3.45E+05 13.79
0.00
gi|344236520 1 61.57 0.02 1.55 40S ribosomal protein S13 [Cricetulus griseus] [MASS=15245] RSP13 1.16E+05 1.79E+05
1.54
gi|354503637 1 61.06 3.85E-03 1.77 PREDICTED: STE20-like serine/threonine-protein kinase isoform 1 [Cricetulus
griseus] [MASS=142185]
SLK 4.87E+04 2.75E+04
0.56
gi|344247607 1 60.45 9.02E-06 20.57 Suprabasin [Cricetulus griseus] [MASS=63932] SBSN 2685.43 5.52E+04
20.56
gi|346421390 1 58.73 0.02 5.36 C-X-C motif chemokine 3 precursor [Cricetulus griseus] CXC3 1.56E+05 2.91E+04
0.19
gi|344250293 1 57.66 0.01 1.86 rRNA 2'-O-methyltransferase fibrillarin [Cricetulus griseus] [MASS=25970] FBL 4.21E+04 7.84E+04
1.86
gi|71043900 1 56.08 0.03 Infinity 60S ribosomal protein L11 [Rattus norvegicus] RPL11 2.89E+05 0
0.00
gi|354468793 1 53.47 1.96E-06 2.49 PREDICTED: leukocyte elastase inhibitor A [Cricetulus griseus] SERPINB 8.27E+04 2.06E+05
2.49
267
Table 3.6: The fold change highlighted in Green or Red denotes the direction in expression of each identified gene in relation to miR-23 sponge clones with
Green being up and Red being down. Tags indicate the implementation of analytical criteria including fold change above 1.5 and an Anova value below 0.01.
gi|344249602 1 51.72 1.48E-03 1.64 Ras GTPase-activating-like protein IQGAP1 [Cricetulus griseus]
[MASS=191377]
IQGAP1 7.88E+04 4.80E+04
0.61
gi|354471687 1 51.7 0.02 1.59 PREDICTED: sorbitol dehydrogenase [Cricetulus griseus] SDH 8.25E+04 1.31E+05
1.59
gi|354504447 1 51.44 1.18E-04 5.77 PREDICTED: hypothetical protein LOC100755523 [Cricetulus griseus]
[MASS=7910]
SPRR2E 1.95E+04 1.13E+05
5.79
gi|344253587 1 50.88 2.66E-03 1.67 Thioredoxin reductase 1, cytoplasmic [Cricetulus griseus] [MASS=61858] TXNDR1 6.61E+04 1.10E+05
1.66
gi|354479983 1 50.38 2.50E-03 1.67 PREDICTED: malate dehydrogenase, cytoplasmic-like [Cricetulus griseus] MDH 4.07E+05 6.79E+05
1.67
gi|354493216 1 48.45 0.03 1.53 PREDICTED: 26S proteasome non-ATPase regulatory subunit 7-like
[Cricetulus griseus]
PSMD7 7.40E+04 4.85E+04
0.66
gi|344254748 1 48.09 2.30E-04 1.62 Cytosolic acyl coenzyme A thioester hydrolase [Cricetulus griseus]
[MASS=35644]
ACOT7 1.44E+05 8.90E+04
0.62
gi|344248399 1 48.08 5.56E-03 1.82 hypothetical protein I79_019528 [Cricetulus griseus] [MASS=271141] PRUNE2 4.81E+04 8.74E+04
1.82
gi|344245117 1 47.43 2.41E-03 1.69 Fermitin family-like 2 [Cricetulus griseus] [MASS=65673] FERMT2 1.95E+05 3.30E+05
1.69
gi|344241413 1 47.22 2.07E-03 1.66 Epidermal growth factor receptor substrate 15-like 1 [Cricetulus griseus]
[MASS=82827]
EPS15L1 3.26E+04 1.96E+04
0.60
gi|344242484 1 46.57 0.01 1.79 Cytoplasmic dynein 1 light intermediate chain 2 [Cricetulus griseus]
[MASS=68816]
DYNC1L12 1.54E+04 2.75E+04
1.79
gi|354499825 1 45.47 0.01 2.33 PREDICTED: heme oxygenase 1-like [Cricetulus griseus] HMOX1 3.45E+04 8.05E+04
2.33
gi|354468396 1 43.47 0.02 1.97 PREDICTED: serine/threonine-protein phosphatase PP1-beta catalytic subunit-
like [Cricetulus griseus] [MASS=37328]
PPP1CB 4.02E+04 7.92E+04
1.97
gi|354495432 1 41.82 5.41E-04 2.24 PREDICTED: asparagine synthetase [glutamine-hydrolyzing]-like isoform 1
[Cricetulus griseus]
ASNS 8.78E+04 3.92E+04
0.45
gi|344244307 1 40.44 5.84E-06 2.77 14-3-3 protein eta [Cricetulus griseus] [MASS=27084] YWHAH 6.50E+04 1.80E+05
2.77
268
We conducted our own quality control tests as a means to identify proteins whose
identification were potentially by chance. The average protein expression of a test group
is compared to the average protein expression of the control group resulting in identifying
DE. By un-supervising the samples, large variation within samples contributing to DE
can be identified. The “6 replicates” of each group (control and test) were stripped of their
identifiers and randomised, ultimately resulting in two groups for comparison whose
occupants potentially spanned samples from both control and test groups. When DE
analysis was carried out in this instance, only two proteins were found to be DE, neither
of which overlapped with those identified in the list of 41 above (Appendix fig. A9).
For example, after carrying out the various comparisons, LETM1 was found to be a
potential target of miR-23 (Fig. 3.85). This figure demonstrates that LETM1 expression
range is tight and consistent throughout control clones (pink box) and significantly
increased in expression in miR-23-depleted clones (blue box). Furthermore, the data
suggests that titrational effects are potentially occurring due to differential sponge
expression, as two triplicate clusters appear to demonstrate distinct patterns of expression
within the miR-23 test samples (Fig. 3.85-Black box).
Figure 3.85: Differential expression of LETM1 in control and miR-23 sponge clones
Figure 3.85: Progenesis software output of the protein expression pattern of LETM1 for each of the
samples replicates within both control (Pink) and test (Blue) groups that ultimately contribute to the
calculation of both fold change and statistical significance.
Furthermore, when both negative control clones were compared, LETM1 was not found
to be DE as indicated in Figure 3.85. Additionally, when both miR-23 clones were
compared, LETM1 was not found to be DE further strengthening that LETM1 is
269
potentially increased as a product of sharing a post-translational controller (miR-23) and
not as a result of clonal variation.
When the control clones (miR-NC sponge clone 2 and 11) were compared, 69 proteins
were found to be DE, compared, to 19 proteins DE between the test clones (miR-23
sponge clone 4 and 6). Interestingly, 4 proteins determined to be DE in miR-23 depleted
clones when compared to controls were also found to be DE between control clones
themselves: Malate dehydrogenase (MDH), heme-oxygenase-1 (HMOX1), non-muscle
caldesmon (CALD1) and 26S proteosome non-ATPase regulatory subunit 7-like
(PSMD7). By comparing both control and test clones against each other it examines the
potential for clonal variation irrespective of the introduction of a miRNA sponge. By
grouping all clones into their respective groups these inter-clonal differences will
ultimately be accounted for resulting in a lower statistical significance. Profiling a larger
panel of clones would be a means to further refine the data output accounting for
expressional variations that are not a product of or indeed specific to the introduction of
a sponge specific for miR-23.
270
Figure 3.86: Progenesis output of 4 proteins DE in control clones
Figure 3.86: Progenesis output diagram of the DE of four proteins in miR-23 clones versus miR-NC
controls. The four proteins listed above, however, were determined to be DE between both control
clones.
3.5.2.2 High Priority targets of miR-23
Of the list of 41 proteins found to be DE across the miR-23 depleted CHO-SEAP clones
when compared to control clones, it was necessary to identify which of these proteins
were DE as a result of miR-23 diversion and not as a byproduct of the CHO cell
phenotype. The online resource, miRWalk (Dweep et al. 2011b), was utilised to generate
a list of potential mRNA targets of miR-23. miRWalk assesses binding potential of a
selected miRNA for all possible mRNAs by identifying a hetpamer complementary
PSMD7
HMOX1
MDH
CALD1
271
sequence to the miRNA seed region and extending or “walking” in the 3’ direction for
further complementarity and scored accordingly. Additionally, miRWalk considers
binding sites within the coding sequences (CDS) and 10kb upstream of the CDS including
the 5’UTR, a criteria not encompassed in the vast majority of prediction algorithms.
Furthermore, miRWalk compares the readouts of up to 8 separate prediction algorithms
(3’UTR only) scoring predicted mRNAs based on the number of algorithms that overlap
in their prediction.
Of the 41 proteins found to be DE, 13 were found to overlap with the miRWalk algorithms
(Table 3.7). These proteins are listed according to their statistical significance (Anova
value) with the addition of the number of algorithms predicting each protein. Given the
nature of the miRNA sponge construct and its influence on the cells translatome, it was
possible to further reduce this list of 13 to a more refined and manageable list of proteins.
As the presence of the miRNA sponge provides a surplus of binding sites for endogenous
miR-23, the activity of this miRNA is being sequestered away from its normal cellular
targets thus ultimately resulting in their translational de-repression. Taking this into
account, DE proteins from this list of 13 that exhibit an increased-fold change would
correlate with the diversion of miR-23 away from its natural suppressive role.
272
Table 3.7: Predicted targets of miR-23 using the miRWalk algorithm
Gene I.D Accession No. FC Anova
Value
Algorithm
overlap
Biological Process
Predicted targets demonstrating up-regulation
LETM1 gi|344244293 2.73 4.33 x 10-7 3 ETS Biogenesis/
Mitochondria
CALD1 gi|344255736 2.16 3.12 x 10-4 1 Calcium Signalling
CAMK2D gi|344255594 1.76 4.95 x 10-4 2 Calcium binding
FERMT2 gi|344245117 1.69 2.41 x 10-3 2 Actin assembly
IDH1 gi|344238611 1.68 2.48 x 10-3 1 TCA cycle / α-KG
TXNRD1 gi|344253587 1.67 2.66 x 10-3 4 Oxidative stress
PPP1CB gi|354468396 1.97 2.01 x 10-2 3 Glycogen
metabolism, protein
synthesis and
chromatin structure
P4HA1 gi|344254200 1.84 1.48 x 10-4 2 Collagen synthesis
Predicted targets demonstrating down-regulation
ABCF1 gi|354487570 1.85 6.64 x 10-5 2 ATP binding
ASNS gi|354495432 2.24 5.41 x 10-4 3 Asparagine
synthesis
IQGAP1 gi|344249602 1.64 1.48 x 10-3 1 Cell morphology
and motility
CXCL3 gi|346421390 5.36 1.95 x 10-2 1 Cell migration and
adhesion
HNRNPA1 gi|4504445 2877.1 4.07 x 10-2 1 pre-mRNA
processing
Of the shortlist of 7 proteins found to be increased in their levels of expression as a result
of miR-23 depletion, several appear to be involved in cellular processes that potentially
reflect the observed phenotype of enhanced mitochondrial function such as LETM1
(mitochondrial biogenesis), IDH1 (TCA cycle, α-ketoglutarate synthesis) and TXNDR1
(oxidative stress response).
Further experiments would be required to validate these proteins as true targets of miR-
23. Western blot could be utilised to determine the sensitivity and accuracy of mass
spectrometry by selecting a small panel of high priority candidates such as LETM1.
Ultimately, functional characterisation of priority targets could be explored by siRNA
273
knockdown of the protein to examine whether their DE could be linked to observed
phenotype.
3.5.2.3 Gene Ontology analysis of differentially expressed proteins
To give an indication if the list of DE expressed proteins within the miR-23 clones
displayed any enrichment for their involvement in similar biological processes, the entire
list of 42 were analysed using the DAVID (Database for Annotation, Visualisation and
integrated Discovery: http://david.abcc.ncifcrf.gov/home.jsp) algorithm. From the
biological processes identified, cellular metabolism was apparent (Table 3.8).
Table 3.8: Gene Ontology of miR-23 sponge DE proteins
GO ID GO Term P-value BH adj.
0015980 Energy derived by oxidation of organic
compounds
8.4 x 10-4 2.5 x 10-1
0006091 Generation of precursor metabolites 1.4 x 10-3 2.2 x 10-1
0051186 Cofactor catabolic process 1.7 x 10-3 1.8 x 10-1
0006006 Glucose metabolic process 2.3 x 10-3 1.8 x 10-1
0055114 Oxidation reduction 3.7 x 10-3 7.3 x 10-1
0006099 Tricarboxylic acid cycle 4.3 x 10-2 7.5 x 10-1
0046356 Acetyl-CoA catabolic process 4.5 x 10-2 7.3 x 10-1
0009060 Aerobic respiration 5.0 x 10-2 7.2 x 10-1
0006084 Acetyl-CoA metabolic process 5.8 x 10-2 7.2 x 10-1
3.5.3 Identification of differential protein expression in isolated mitochondria
miR-23 depletion was previously observed to impact on mitochondrial function in
addition to mitochondrial-associated proteins being identified as increased such as
LETM1. Mitochondrial extraction was carried out 72 h into culture as before (see section
2.5.3 of Materials and Methods) and processed for label-free mass spectrometry
following the same procedure as previously described for proteomic profiling of whole
cell cultures.
274
3.5.3.1 Differentially expressed proteins
When label-free mass spectrometry was carried out on the isolated mitochondria of the
same miR-23 sponge clones (4 and 6) and miR-NC sponge clones (2 and 11), 499 proteins
were determined to be DE exhibiting a fold change of ≥ 1.2 and a peptide identification
of ≥ 1 (Appendix fig. A10). When this list was subjected to more stringent analytical
filters (≥ 1.5 FC in addition to ≥ 2 peptide identifications), a shorter list of 99 proteins
were found to be DE. Overlap of the entire list of 499 with the 42 found to be DE in the
whole cell identified 15 proteins common to both lists (Table 3.9) demonstrating the same
direction of expression in addition to a similar FC in most cases (10 up-regulated and 5
down-regulated). One interesting observation was the identification of the cytoplasmic
MDH1 to be upregulated by 1.79-fold in miR-23 sponge clones in the whole cell lysate.
Whereas the mitochondrial-specific MDH2 and the cytoplasmic MDH1 were identified
to be up-regulated by 1.3 and 2.7, respectively, in mitochondrial enriched miR-23 sponge
samples.
Table 3.9: DE proteins common to both Whole cell and Mitochondrial
Whole Cell Lysate Mitochondria
Protein FC Protein FC
GSTM6 2.73 GSTM6 2.4
MFGE8 2.49 MFGE8 2.2
CALD1 2.77 CALD1 1.9
P4HA1 2.24 P4HA1 1.7
GAA 1.54 GAA 2.5
RPS13 1.66 RPS13 1.3
SBSN 1.68 SBSN 3.6
TXNRD1 1.86 TXNRD1 1.8
HMOX1 Infinity HMOX1 2
MDH1 1.79 MDH1 2.7
YWHAH 2877.10 YWHAH 1.7
EPRS 2.49 EPRS 2.6
ABCF1 20.57 ABCF1 2.4
GAPDH 1.74 GAPDH 1.4
HNRNPA1 1.76 HNRNPA1 1.5
IQGAP1 1.82 IQGAP1 1.6 Table 3.9: DE proteins common to both label-free MS list (Whole cell lysates and Mitochondrial
enriched). Fold-change (FC) comparisons are listed to demonstrate similarities across both lists and
MS sensitivity in addition to an identical direction of expression (Green – upregulated and Red –
down regulated). FC and colour-coding refers to the expression in miR-23 depleted clones compared
to controls.
275
When cell component association of both DE protein lists from the whole cell (42
proteins) and the mitochondrial enriched (499 proteins) was assessed, there was a 10%
enrichment for proteins associated with the mitochondria (47 proteins) in the
mitochondrial enriched samples compared to no apparent mitochondrial association in
the case of the DE proteins from the whole cell lysate. This indicated a significant
enrichment over what would be expected by chance.
3.5.3.2: Gene Ontology analysis
Similar to the case of miR-23 sponge whole cell lysate DE proteins, GO analysis was
performed for the 499 proteins DE in the mitochondrial enriched samples. Biological
processes associated with translation, RNA processing, ribosome biogenesis and protein
folding were found to be enriched. This is potentially in agreement with the miR-23
depleted associated phenotype of high SEAP production. When all ribosomal proteins
found to be DE were further analysed, it was found that 26 out of 35 (74.2%) were up-
regulated in miR-23 sponge clones, in keeping with what would be expected in cells with
an increased protein output. Only one of these ribosomal proteins was determined to be
specific to the mitochondria (Mitochondrial Ribosomal protein S7 – MRPS7) and was
decreased in its expression.
With a far greater degree of statistical significance, GO on the 499 DE proteins also
indicated biological processes related to cellular respiration including aerobic respiration
and the Tricarboxylic acid cycle (Table 3.10).
276
Table 3.10: GO for 499 DE proteins
GO ID GO Term P-value BH adj.
0006414 Translation elongation 1.8 x 10-34 3.7 x 10-31
0006412 Translation 2.9 x 10-26 3.0 x 10-23
0006396 RNA processing 2.4 x 10-12 1.6 x 10-9
0006397 mRNA processing 1.4 x 10-10 5.7 x 10-8
0016071 mRNA metabolic process 1.4 x 10-9 4.9 x 10-7
0006457 Protein folding 5.6 x 10-9 1.7 x 10-6
0009060 Aerobic respiration 3.9 x 10-7 7.3 x 10-5
0046356 Acetyl-CoA catabolic process 2.3 x 10-6 3.7 x 10-4
0006099 Tricarboxylic acid cycle 2.3 x 10-6 3.6 x 10-4
0042254 Ribosome biogenesis 8.9 x 10-6 1.1 x 10-3
0006084 Acetyl-CoA metabolic process 2.0 x 10-5 2.1 x 10-3 Table 3.10: GO categories for DE expressed proteins in mitochondrial enriched miR-23 samples
versus miR-NC clones using DAVID.
Finally, of the 47 DE proteins that were found to be associated with the mitochondrion
when assessed for cell compartment affiliation, gene ontology enrichment was carried out
as a means to determine which elements of mitochondrial function were predominately
under regulation (Table 3.11). The highest scoring biological processes found to be
enriched were aerobic respiration, cellular respiration, Tricarboxylic acid cycle and
acetyl-CoA catabolic processes indicating that, of the mitochondrial related proteins that
were DE, a significant proportion of them were influencing pathways responsible for
cellular energy, as indicated by the observation of enhanced oxygen consumption
(OXPHOS) by miR-23 depleted clones (Section 3.4.6).
277
Table 3.11: GO analysis of DE proteins associated with the mitochondria
GO ID GO Term P-value BH adj.
0009060 Aerobic respiration 1.2 x 10-13 7.8 x 10-11
0045333 Cellular respiration 1.6 x 10-11 5.3 x 10-9
0006099 Tricarboxylic acid cycle 7.0 x 10-11 1.6 x 10-8
0046356 Acetyl-CoA catabolic process 7.0 x 10-11 1.6 x 10-8
0009109 Coenzyme catabolic process 1.6 x 10-10 2.7 x 10-8
0006084 Acetyl-CoA catabolic process 5.0 x 10-10 6.7 x 10-8
0051186 Cofactor catabolic process 5.0 x 10-10 6.7 x 10-8
0015980 Energy derivation by oxidation of organic
compounds
5.7 x 10-10 6.4 x 10-8
0006732 Coenzyme metabolic process 9.7 x 10-10 9.4 x 10-8
0055114 Oxidation reduction 5.2 x 10-9 4.4 x 10-7
0006091 Generation of precursor metabolites 3.7 x 10-8 2.5 x 10-6
0043648 Dicarboxylic acid metabolic process 3.8 x 10-8 2.4 x 10-4
0006105 Succinate metabolic process 3.5 x 10-4 2.0 x 10-2
0022904 Respiratory electron transport chain 1.1 x 10-3 5.5 x 10-2
0022900 Electron transport chain 5.6 x 10-3 2.4 x 10-1
0007005 Mitochondrion organisation 9.5 x 10-3 9.5 x 10-1
Table 3.11: Gene Ontology categories for differentially expressed proteins in mitochondrial enriched
miR-23 samples versus miR-NC clones using DAVID.
However, upon further interrogation of the proteins involved in these biological processes
associated with mitochondrial energy production, several participating candidates were
found to be decreased in their levels of expression such as IDH3A (Isocitrate
dehydrogenase [NAD] subunit alpha) and COX5A (Cytochrome c oxidase subunit 5A).
Further analysis of the proteins related to these processes will be considered later within
the discussion (Section 4.8).
3.5.3.3: Potential targets of miR-23
Similar to the exercise carried out in the case of the list of DE proteins from the whole
cell lysate experiment, the predicted output of targets of miR-23 was overlapped with the
499 DE proteins in the mitochondrial proteomics profiling experiment. This generated a
total list of 143 potential targets of miR-23 (64 up-regulated and 79 down-regulated,
Table 3.11). To further refine this extensive list of potential targets, those that appeared
in the short list of 99 DE proteins reduced this list to 32 (10 up-regulated and 22 down-
regulated), highlighted in yellow in Table 3.12. Finally, as it would be expected that the
diversion of miR-23 away from its endogenous targets should result in the translational
278
de-repression of potential targets, only the 10 upregulated proteins from this short list
were considered high priority.
Table 3.12: Potential targets of miR-23 from isolated mitochondria
Gene ID FC Anova (p) Algorithm Overlap
Genes Upregulated in miR-23 sponge clones
LIMA1 1.7 1.41E-06 3
CALD1 1.9 2.30E-06 2
HMGB2 1.5 5.57E-06 3
AKAP12 6.9 9.29E-06 4
SCP2 1.6 1.06E-05 1
P4HA1 1.7 1.97E-05 2
CALU 2 2.13E-05 2
CAPRIN1 1.8 2.25E-05 1
ITGB1 1.6 2.66E-05 2
CTSB 1.7 2.70E-05 1
TXNRD1 1.8 2.92E-05 1
CLINT1 2 4.73E-05 1
HSP90B1 1.5 4.78E-05 3
ALCAM 3.1 7.11E-05 3
TWSG1 2.5 8.00E-05 3
CNN2 1.7 8.38E-05 3
NUFIP2 1.4 0.00034 4
PSAP 3.5 0.00041 1
MAP4 1.4 0.00044 3
ANXA2 1.4 0.00051 2
BCL9L 2.3 0.00074 2
CETN3 1.9 0.00076 2
HDDC2 1.7 0.00112 3
TPI 1.2 0.00127 2
WBP2 1.3 0.00172 4
PRKCSH 1.4 0.00181 3
AIM2 2 0.00191 2
SNX1 1.5 0.00217 2
MAGOH 1.4 0.00231 2
LAMC1 1.5 0.00381 1
LAMP1 2.9 0.0042 4
SET 1.6 0.00451 2
IDI1 2 0.00466 1
PDIA6 1.8 0.00534 3
CACYBP 1.3 0.00543 3
TMED9 1.4 0.00601 1
HEXIM1 1.3 0.00685 4
CRK 1.9 0.00926 2
PLEKHO2 1.4 0.00936 3
RAB14 1.4 0.00956 3
279
PPIB 1.4 0.0098 1
RPL10 1.7 0.01005 4
BCLAF1 2.2 0.01103 4
CTSL 1.3 0.0118 1
SNAP29 1.4 0.01615 2
FXR1 1.4 0.01785 1
LMAN2 1.1 0.01836 3
RRAS2 1.5 0.01912 4
PEX19 1.3 0.01968 2
RPL27A 1.7 0.02155 3
TPI 1.2 0.02234 2
FKBP10 1.4 0.02374 1
SEPT7 1.2 0.02385 3
RCN2 1.6 0.02611 1
HYPK 2.2 0.0274 2
DSC1 1.2 0.02788 2
PRPF40A 1.5 0.03643 3
FUBP1 1.5 0.03698 2
RPL31 1.5 0.03785 1
TMOD3 1.2 0.0386 4
RPS23 1.5 0.04037 2
ENO2 1.2 0.04215 4
GIGYF2 1.1 0.04946 2
PRDX2 1.3 0.04984 1
Genes Down-regulated in miR-23 sponge clones
IQGAP1 1.6 3.40E-05 1
SKIV2L2 2.1 5.12E-05 3
TOP2A 1.7 5.99E-05 2
VCP 1.9 6.71E-05 3
VCL 1.5 7.35E-05 1
MYH10 1.5 8.17E-05 2
PLAA 1.8 0.00011 2
GRSF1 1.4 0.00012 3
H2AFV 3 0.00013 2
FASN 2.8 0.00014 2
SLC7A1 2 0.00026 3
RRP15 1.6 0.00027 4
HDAC2 1.7 0.00028 3
MCM2 1.8 0.00032 2
OXCT1 1.5 0.00041 3
SND1 1.9 0.00041 2
WDR43 1.8 0.00042 2
NPM1 1.3 0.00048 1
HSPH1 2.3 0.00053 2
RRM1 1.3 0.00056 1
ABCF1 2.4 0.00057 2
CLTC 1.7 0.00062 2
LRPPRC 1.6 0.00069 3
280
FLNB 1.5 0.00073 2
SLC25A3 1.8 0.00083 1
LYAR 1.3 0.00088 1
CAND1 1.9 0.0009 1
TES 1.4 0.00104 2
EIF4G1 1.4 0.00108 1
TMPO 1.3 0.0011 3
HMGCS1 1.4 0.00121 2
MAP1B 1.5 0.00165 1
PPP2R1A 1.3 0.00195 2
MATR3 1.6 0.00196 1
SMARCC1 1.3 0.00199 3
MRPS7 1.4 0.00229 1
SUCLA2 1.4 0.0023 1
HAT1 1.5 0.00233 1
TOP1 1.5 0.00269 4
RPN1 2.3 0.00309 1
GATAD2B 1.4 0.00327 1
SF3B3 2 0.00328 1
SFPQ 1.3 0.00379 2
RBM14 1.4 0.00379 1
DDB1 1.5 0.00511 2
MKI67 1.5 0.00516 2
WNK1 1.3 0.00519 4
SPTBN1 2 0.00575 2
HNRNPU 1.8 0.0058 3
DDX21 1.8 0.00583 2
HNRNPC 3.1 0.00664 1
MTAP 1.5 0.00757 3
CPSF6 1.4 0.00803 3
ILF3 1.4 0.00914 1
HNRNPA1 1.5 0.01038 1
KIF5B 1.3 0.01132 1
SMC1A 1.5 0.01164 4
DDX5 1.4 0.0128 2
ZC3HAV1 1.5 0.01293 2
CTNND1 1.4 0.01374 2
HIST1H2AG 2.1 0.0142 1
LBR 1.6 0.01466 2
SPTBN1 1.8 0.01533 3
ACTR2 1.2 0.01651 3
TBL2 1.5 0.01816 4
HNRNPA2B1 1.2 0.0182 3
HIST1H2AG 1.4 0.01858 1
OPA1 1.3 0.01882 2
KHSRP 1.6 0.02496 3
SAFB 1.3 0.02618 3
CSE1L 1.5 0.02706 4
281
RRP12 1.5 0.0278 1
EZR 1.5 0.02814 2
VAPA 1.2 0.02995 4
IRF2BP2 1.2 0.03608 2
NCAM1 9.4 0.03793 3
CKAP4 1.2 0.0392 3
IQCE 1.3 0.04587 3
MAPRE1 16.6 0.0487 4 Table 3.12: Potential targets of miR-23 are shown by overlapping the hits generated by the miRWalk
algorithm with the list of DE proteins from isolated mitochondria. The list is broken up into two, low
priority targets found to be down-regulated in miR-23 clones and high priority targets found to be
up-regulated in miR-23 clones. Cells highlighted in yellow passed more stringent criteria of over 1.5-
fold change in expression with the addition of 2 or more peptide identifications.
Of the two short-lists generated from both label-free MS experiments (DE proteins that
overlapped with prediction algorithms demonstrating an increased expression) and taking
into account the two common targets CALD1 and P4HA1, a list of 18 targets were
identified to be high priority targets of miR-23 that would be subject to further validation.
These targets are LIMA1, HMGB2, SCP2, CALU, ITGB1, CTSB, SNX1 and PDIA6 in
the mitochondrial extracted list and LETM1, CaMKIIδ, FERMT2, IDH1, TXNDR1 and
PPP1CB in the whole cell lysate.
282
Chapter 4
Discussion
283
4.1 CHO cell engineering using miR-23
miR-23 depletion presented itself as the most viable candidate for stable miRNA
intervention when assessed in CHO-K1-SEAP cells. Although not directly derived from
the initial miRNA profile carried out on CHO cells with varying growth rates, previous
manipulation of the miR-23~27~24 cluster using miRNA sponge technology proved to
be an attractive engineering route enhancing both CHO-SEAP cell growth and product
yield.
Stable depletion of a single component of this cluster, miR-23, in the same CHO-SEAP
cell line was observed not to play a role in growth but did elicit an improved product yield
of ~ 3-fold. These high producing CHO clones were identified to possess an enhanced
mitochondrial function with an increase in oxidative metabolism of ~ 30% centred on
complex I and II of the mitochondrial electron transport chain. Protein processing and
biogenesis has been identified as an energetically demanding process highlighting the
potential contribution of enhanced mitochondrial activity to accommodate this cellular
burden. Calcium signalling not only is suggested to be an ATP sensor within the
endoplasmic reticulum (Kaufman, Malhotra 2014) thus monitoring energy requirements
but is also implicated in TCA cycle function through the activation of vital enzymes of
the TCA cycle thus driving energy provisions (Osellame, Blacker & Duchen 2012). It
was discovered that the increase in SEAP yield was potentially attributed partially to an
increase in SEAP transcription. This possess the question of whether there lies a
reciprocal relationship between transcriptional activity/protein processing with enhanced
OXPHOS or vice versa.
Label-free mass spectrometry identified various protein targets whose expression was
increased suggesting potential targets of miR-23 itself that are associated with
mitochondrial function and cellular metabolism, IDH1/MDH1 and LETM1, in addition to
proteins implicated in transcriptional control, CaMKIIδ and 14-3-3-η (Fig. 4.1). This
schematic forms the basis of our hypothesis of the impact miR-23 is having in these CHO-
SEAP clones ultimately boosting productivity of the bioprocess.
284
Figure 4.1: Schematic of miR-23s hypothetical role in rCHO-SEAP cells
Figure 4.1: The depletion of miR-23 could impact directly or indirectly on mitochondrial function given the suggested panel of potential targets from the
proteomics profiling. Isocitrate dehydrogenase (IDH1) and Malate dehydrogenase (MDH1) provide TCA cycle metabolites. LETM1 is an inner mitochondrial
membrane (IMM) protein acting as a Ca2+/H+ antiporter with calcium activating TCA cycle enzymes in the matrix while H+ provides energy to the proton
gradient. Indirectly, increased expression of CaMKIIδ targets HDAC4 for phosphorylation, tagging it for 14-3-3-η thus preventing nuclear localisation. HDAC4
inhibits transcription through deacetylation. Enhanced SEAP processing in the endoplasmic reticulum (ER) is an energy-demanding process, releasing Ca2+ as
ATP is used which stimulates mitochondrial oxidative metabolism.
285
4.1.1 Selection of miR-23 for stable depletion
The pioneering study by Gammell et al., identified a panel of miRNAs to be DE upon
hypothermic growth conditions such as let-7f and miR-21 (Gammell et al. 2007). Two
cluster components, miR-27a and miR-24, were found to be up-regulated upon
temperature shift-induced growth arrest and late culture growth arrest. Despite being
transcribed from a single transcriptional unit, the levels of miR-23a remained unchanged,
an outcome influenced by RNA-binding proteins and processing pathway (Chhabra et al.
2009). Interestingly, these miRNAs demonstrated no change in expression when
evaluated in CHO cells with varying growth rates (Clarke et al. 2012). Previous
characterisation within our research group by Dr. Noelia Sanchez demonstrated that
transient overexpression of miR-23a in CHO-K1-SEAP cells elicited a growth arrest of
61% whereas its inhibition did not reverse the phenotype. Transient overexpression of
miR-24 induced a 27% arrest in cell numbers while transient inhibition reversed this
phenotype to improve cellular growth by the same magnitude. miR-27a overexpression
in two individual cell lines (CHO-1.14 and CHO2B6) resulted in a growth reduction and
improvement, respectively, demonstrating a cell specific influence. It was for this reason
that targeting of the entire miR-23a~27a~24-2 cluster using miRNA-sponge technology
was explored resulting in a 1.7-fold increase in cell numbers and a 2-fold increase in
specific productivity.
This was a beneficial observation in that specific productivity is often observed to be
compromised during enhanced growth (Sunley, Tharmalingam & Butler 2008a). The co-
existence of a fast growing CHO cell and the retention of increased specific productivity
in the case of miR-23a~27a~24-2 depletion offers encouragement in terms of a viable and
robust engineering strategy. CHO cell engineering using decoy technology against miR-
7 (Sanchez et al. 2013b) or stable miR-17 overexpression (Jadhav et al. 2014) have both
enhanced CHO cell growth without the compromise of specific productivity achieving 2-
3 fold improvements in product yield.
miR-23 has also been associated with the reprogramming of cellular bioenergetics (Gao
et al. 2012). The “Warburg effect” describes the propensity of cancer cells to avidly take
up glucose and almost exclusively convert it to lactate in a process known as aerobic
glycolysis (Vander Heiden, Cantley & Thompson 2009). Although energetically
inefficient, it provides critical fuel for biomass production (Catapleurosis) and it
highlights the role that energy metabolism plays in tumorigenesis (Fig. 4.2).
286
It has however been demonstrated that weakening of the TCA cycle by Acetyl-CoA
starvation mediated through the Warburg effect (Dell'Antone 2012) can be overcome by
anapleurotic reactions that resupply mitochondrial intermediates of the TCA cycle such
as α-ketoglutarate through miR-23s role in glutaminolysis (Gao et al. 2012).
Figure 4.2: Schematic of Aerobic and Anaerobic glycolysis in various cellular
conditions
Figure 4.2: A schematic representation of “normal” channelling between Oxidative phosphorylation
and anaerobic glycolysis under high and low oxygen concentrations, respectively. In addition, the
utilisation of excess glucose sources via the Warburg effect in both normal and diseased tissues. This
figure was sourced from (Vander Heiden, Cantley & Thompson 2009).
The primary carbon sources, Glucose and Glutamine, have been extensively considered
in the area CHO cell research. Monoclonal antibody production has been positively
correlated with the metabolism of glucose and lactate (Chen et al. 2012). Metabolic by-
product accumulation such as lactate and ammonia from glucose and glutamine
metabolism, respectively, has been extensively considered for their impact on cell growth
and productivity, not to mention product quality (Yang, Butler 2000). Studies centred on
glutamate supplementation have contributed to enhanced productivity and
galactosylation of recombinant IgG (Hong, Cho & Yoon 2010) and optimisation of fed-
287
batch cultures through galactose/glutamate replacement (Altamirano et al. 2000). Various
CHO cell phenotypes have also been affiliated with particular metabolic states such as
growth, glycolytic metabolism, productivity and oxidative metabolism (Templeton et al.
2013). Metabolic re-programming using miR-23 depletion could enhance productivity
through enhanced mitochondrial function while maintaining growth through an
unperturbed glycolytic process.
4.1.2 GFP as an effective reporter for clone selection and sponge efficiency
The utilisation of a reporter construct has enabled the selection of rCHO clones
demonstrating high specific productivities (Lai, Yang & Ng 2013a). Additionally, they
have proved useful in the implementation of miRNA-based technology, such as miR-
sponges. We utilised a destabilised GFP (deGFP) reporter construct to monitor miRNA
sponge efficacy (Kitsera, Khobta & Epe 2007). When expressed at supraphysiological
levels, miRNA sponges act as decoy transcripts (Mullokandov et al. 2012) but when
expressed at physiological levels they can be used as miRNA “sensors” allowing for the
evaluation of endogenous miRNA activity (Mansfield et al. 2004). Despite a wide range
of GFP intensities, we observed an average reduction in GFP expression across the miR-
23 sponge clonal panel when compared to the miR-NC controls. This reduced GFP
expression was presumably a result of endogenously expressed miR-23a/b, which should
not influence the non-specific sponge, as observed by Sanchez and colleagues for miR-7
(Sanchez et al. 2013b).
To further confirm the specificity of the miR-23 sponge construct, mixed pools were
transfected with a miR-23b mimic as a means to repress translation of the GFP reporter.
One interesting observation however was that transient transfection of the miR-23b
mimic resulted in a mild repression of GFP (Fig. 3.31 B). One explanation is the reported
low proficiency of miR-23 derived from its “weak seed”-pairing stability (SPS) due to its
A+U-rich seed region (Garcia et al. 2011). Furthermore, miRNAs with A+U-rich seeds
have a greater target abundance due to the A+U-rich composition of 3’UTRs (Grimson
et al. 2007) ultimately titrating miRNA function (Franco-Zorrilla et al. 2007). With this
in mind, we attempted to transfect a miRNA that was specifically designed to mimic the
natural structure of the miR-23b duplex (pre-miR-23b*-mod) meaning that upon
exogenous introduction into cells, would be processed based on thermodynamic stability
of the duplex ends giving rise to both miR-23b and miR-23b*. When transfected, both
288
miR-23 sponge mixed populations and two clones (4 and 6), demonstrated a 90%
reduction in GFP (Fig. 3.31 Bpre-miR-23b*-mod and Appendix Fig. A14). These
observations suggested that the first miR-23b mimic was not being processed efficiently
whereas the modified, pre-miR-23b* mod was, achieving GFP inhibition. miR-23 sponge
specificity was also explored by assessing the endogenous levels of mature miR-23b in
clones stably expressing the sponge construct compared to controls. A 5-fold reduction
in mature miR-23b levels were observed in our clones when compared to clones
expressing a non-specific sponge (Fig. 3.67).
The rapid degradation of mature miRNAs have been reported in cases of target
availability such as miR-27a/b by murine cytomegalovirus (Marcinowski et al. 2012),
miR-15a/16-1 by hepatitis B virus (HBV) (Liu et al. 2013), miR-122 by HBV (Li et al.
2013c) and the miR-let-7 family have been observed using a miRNA sponge (Yang et al.
2012). Although contradictory observations have been reported for miR-7 (Sanchez et al.
2013b), the mechanism behind the decrease of endogenous miR-23 in stable sponge
clones remains unceratin.
4.1.3 Stable depletion of miR-23 improves CHO-SEAP productivity
Specific productivity was enhanced by 30-50% in the case of CHO mixed pools with an
average 3-fold improved SEAP activity in a clonal panel. Of the four top SEAP producing
clones selected, 3 demonstrated the highest level of GFP intensity compared to the
majority of the miR-23 clones suggesting the highest levels of miR-23 depletion. One
clone in particular however, miR-23 clone 2, possessed the third lowest GFP expression
yet was the fourth highest producing clone. This would suggest that the intensity of the
fluorescent signal is not necessarily proportional to the level of productivity. miR-23
sponge clone 9 for example displayed a high GFP signal compared to miR-23 sponge
clone 7 (Fig. 3.39) yet its productivity value is one of the lowest. It is possible that by
selecting clones with mid-high GFP intensities, a titrational influence of miR-23 depletion
is lost and the productivity differences are due to clonal variation. The selection of clones
exhibiting a low-medium GFP intensity might allow the assessment of a correlation of
sponge expression with productivity improvements.
Protein folding and secretory bottlenecks have been shown to restrict mammalian cells
from exploiting their physiologic production capacity (Tigges, Fussenegger 2006). As a
means to investigate whether miR-23 depleted clones possessed an enhanced ability to
289
process and secrete the SEAP protein, we determined there to be a ~2-fold increase in the
levels of the SEAP mRNA transcript when compared to control clones. This indicated
that there were no processing bottlenecks in either control or test clones as the levels of
SEAP transcript reflected SEAP enzymatic activity.
mRNA stability has been demonstrated to contribute to specific productivity as in the
case of CHO cells expressing TNFR-Fc subject to hypothermic growth conditions (Kou
et al. 2011). After translational inhibition through Actinomycin-D treatment, the rate of
mRNA decay was determined by calculating the slope across the various time points
when qPCR was carried out. Ideally, a similar decrease at each time when compared to
time zero would provide information as to the rate of decay independent of varying Ct
values due to differing transcript concentrations. It was observed that the endogenous
control (β-actin) displayed an accelerated rate of decay when compared to the SEAP
transcript, therefore 2-ΔΔCt calculation would not have been suitable. miR-23 sponge
clones demonstrated no change in decay suggesting there to be no increase in SEAP
mRNA stability when compared to controls.
It has been strongly suggested that lessons from nature could aid in the identification of
attractive engineering targets for the use in CHO cell manipulation, in particular the
antibody producing B-cell (Dinnis, James 2005). Resting B-cells, once stimulated,
differentiate into nature’s own antibody factories, plasma cells, and have presented
engineering targets such as XBP-1 (Tigges, Fussenegger 2006), BiP and PDI (Pybus et
al. 2013) and ATF4 (Ohya et al. 2008b). In the case of miR-23, its down-regulation has
been associated with the progression of chronic lymphocytic leukemia (Chhabra, Dubey
& Saini 2010b). A recent study by Li and colleagues demonstrated that malignant
transformation of B-cells closely resemble the activation of resting B-cells (Li et al.
2011a). Among the panel of miRNAs found to be similarly deregulated across both
phenotypes, miR-23a/b, miR-24 and miR-27b were shown to be down-regulated in
activated B-cells. Additionally, hematopoietic progenitors expressing the miR-23a cluster
and cultured in the presence of B-cell promoting conditions were observed to undergo a
dramatic reduction in B-lymphopoeisis (Kong et al. 2010). The expression of miR-23a
has also been observed to be negatively associated with the up-regulation of Blimp-1 (B
lymphocyte-induced maturation protein 1) (Parlato et al. 2013). These studies
demonstrate that the miR-23a cluster plays a central role not only in regulating the
molecular frame work of B-cells but also further in their activation. B-cells are of an equal
290
size to that of CHO cells (10-15 µm) with a specific production capability of 100
pg/cell/day (Dinnis, James 2005), the suggested maximum production capacity of CHO
(De Jesus, Wurm 2011). As B-cells are considered antibody factories, the implication of
miR-23 and its cluster components in the activation of B-cells could suggest a similar
influence in CHO-SEAP cells, specifically stimulating productivity in the case of miR-
23.
4.1.4 miR-23 depletion improves CHO cell growth in a cell type and product-
dependent manner
Although miR-23 depletion was not observed to impact on the growth of rCHO-SEAP
cells, it did appear to elicit a positive impact on a second cell line, rCHO-1.14 upon its
stable depletion, both derived from the same CHO-K1 parent. In both stable mixed
populations and clonal populations cultured in a 1 mL format, there was an increase in
cell growth to the detriment of IgG productivity. Various expression patterns of the
members of miR-23 have been recorded across many cancer states with over-expression
dominating (Table 4.1). Inhibition in this instance enhanced growth within the CHO-1.14
cell line. It is possible that this cell line exhibits a completely different set of gene targets
as its inherent growth is markedly reduced when compared to CHO-SEAP cells. The
oncogenic transcription factor, c-myc, despite being demonstrated to inhibit the
expression of miR-23a/b to ultimately alleviate their inhibitory role on glutaminase (Gao
et al. 2009), has also been shown to up-regulate the miR-23a cluster in mammary
carcinoma (Li et al. 2013).
Furthermore, a significant improvement in apoptosis resistance was mediated by the
depletion of miR-23 in CHO-1.14s in batch culture. miR-23a has been demonstrated to
directly inhibit the caspase-inhibitor, X-linked inhibitor of apoptosis (XIAP) (Siegel et al.
2011). XIAP over-expression has been previously explored in CHO cells exposed to
sodium butyrate treatment as a means to prolong cell survival while exploiting the
induced high specific productivity phase (Kim, Kim & Lee 2009) offering no viability
benefit. A prior study demonstrated that stable XIAP over-expression could inhibit CHO
cell apoptosis upon Sindbis virus infection (Sauerwald, Betenbaugh & Oyler 2002b). This
suggests that the benefit of XIAP can be both cell line specific and potentially influenced
by the route of apoptosis activation. As a result, this could be why miR-23 depletion offers
a protective phenotype in rCHO-1.14 cells but not in rCHO-SEAP cells under batch
291
conditions. Ongoing work by a colleague within our research group has demonstrated that
miR-23 retains its functionality in targeting XIAP in CHO cells thus validating its capacity
to regulate the gene but again in a cell type and context-dependent manner.
It was observed that the improved growth phenotype came hand-in-hand with a reduction
in specific IgG productivity. However, upon clonal characterisation, there was a mixture
in the observed phenotype. Although all clones stably depleted of miR-23 retained a fast
growing phenotype with increased maximal cell density, volumetric productivity was
enhanced in the case of 2 out of the three clones. This could have been due to the bias
introduced during clonal selection, when clones were selected based on their high
volumetric productivity in addition to growth. Furthermore, a single control clone
demonstrated an enhanced viable phenotype exposing the natural variation in clones. It is
possible that the influence on rCHO-SEAP cells in regards to growth is a factor of the
expressed mRNA targets when compared to the rCHO-1.14. Inducing a faster growth
phenotype may not be possible due to the limited availability of miR-23s targets in an
already fast growing cell type.
292
Table 4.1: Table of expression of miR-23 cluster components in various disease
states
Table 4.1: Deregulated expression of the miR-23 cluster paralogs in various disease states us
demonstrated above. “u” denotes up regulation whereas “d” denotes down-regulation. The table was
sourced from the review (Chhabra, Dubey & Saini 2010b).
4.1.5 Clonal variation and culture conditions impact on the CHO cell phenotype
The process of recombinant cell line generation has its inherent pitfalls depending on the
context. For example, non-site specific integration of the experimental transgene, in our
case the miR-23/NC sponge constructs, can result in various genetic phenotypes such as
copy number variation, intensity of expression due to insert location and potential
293
insertional mutagenesis. Random integration and amplification may also interfere with or
de-regulate endogenous genes (Winnard, Glackin & Raman 2006) ultimately creating the
potential for variation in other cellular traits (Kim, Lee 2007). This in particular was
observed when cellular proliferation was characterised upon the isolation of a panel of
clones derived by FACS from the original CHO-SEAP mixed pool. Across both miR-23
and miR-NC sponge clones, a range of growth profiles were identified despite the
observation of no gross differences in growth of the recombinant mixed pool populations.
This heterogeneity within clones, as a function of the transgenesis process, is routinely
exploited in the selection of clones showing an advantage for a particular phenotype
(Pichler et al. 2011).
The parental rCHO-SEAP cell used as the basis for stable miR-sponge generation was a
clonal line and presumed to demonstrate a similar level of SEAP production and genetic
homogeneity. An additional layer of complexity has been demonstrated in that not only
are there variations in the performance from clone to clone but that protein expression
can be heterogeneous within clones themselves. One study demonstrated a 50-70%
deviation in antibody expression within clones upon 2-11 generations (Pilbrough, Munro
& Gray 2009). This could have potentially contributed to a production variance,
originating within the parental CHO-K1-SEAP, across all clones generated irrespective
of the process of recombinant cell line generation.
One very interesting observation was that when miR-23 sponge clones were selected and
cultured in a 1 mL suspension format, the maximum achievable cell density on day 4 was
less than half of that of miR-NC sponge clones, again in contrast to the similar growth
profiles observed in the case of mixed pools. The similarity in cell numbers between both
clonal panels on day 2 of growth suggested that miR-23 depletion itself was not exerting
a direct impact on the CHO cell growth. Interestingly, of the 4 clones selected from this
panel demonstrating a range of maximal cell densities (~1 x 106 to ~4 x 106 cells/mL in
1 mL), each clone exhibited a similar maximal cell density of ~5 x 106 cells/mL upon
culture in a 5mL volume in a vented spin tube. Cell line behaviour has been previously
shown to be sensitive to vessel type and process conditions such as pH, temperature and
dissolved oxygen tension (Trummer et al. 2006, Liu et al. 2011). A comprehensive study
by Porter and colleagues (Porter et al. 2010) followed the progression of 175 antibody-
secreting clones isolated from the same transfected population throughout various phases
of cell line characterisation using a “typical strategy” of cell line development (Fig. 4.3).
294
It was evident that clonal performance was dependent on the culture environment which
ultimately presents a difficult task when empirically selecting clones for phase
development.
Early stage platforms tend not to have strict regulation on particular environmental
conditions such as pH and dissolved oxygen tension (DOT). In the case of our miR-23/NC
sponge screen, the microenvironment of the parafilm-sealed 24-well suspension plate
potentially possessed a less consistent oxygen environment as opposed to the vented spin
flasks, potentially contributing to the miR-23 sponge specific growth inhibition.
Kulshreshtha and colleagues demonstrated a panel of miRNAs responsive to hypoxic
growth conditions with miR-23/-27 and -24 being upregulated (Kulshreshtha et al. 2007).
An independent study demonstrated a reduced expression of miR-23a/b and miR-27a/b
in hypoxia induced apoptosis with an anti-correlation with Apaf-1 (Chen et al. 2014).
Overexpression of these two miRNAs was further demonstrated to protect against
apoptosis. Both of these studies suggest a protective role of miR-23 in response to
hypoxia. It is possible that depletion of miR-23 in CHO-SEAP clones is preventing these
CHO clones from responding to the potential hypoxic conditions of this culture format
compared to control clones resulting in the reduced cell viability observed.
Interestingly, despite the reduced maximal cell densities, a similar average 3-fold increase
in volumetric SEAP productivity was exhibited when compared to individual clones
selected for 5 mL culture. This retention in volumetric SEAP activity appeared to be as a
result of enhanced specific productivity in the case of clones cultured at 1 mL volume
potentially attributed to a prolonged residence in the highly transcriptionally active phase
of the cell cycle, G1 (Kumar, Gammell & Clynes 2007).
295
Figure 4.3: CHO clone behaviour across multiple phase development platforms
Figure 4.3: Heterogeneity within clonal populations along with the variations in performance across
multiple culture conditions ultimately contributing to the complexity in clonal selection during phase
development. A) Represents the “selected top 10” based on their volumetric productivity in batch
suspension culture and B) represents the “real top 10” based on their volumetric productivity in fed-
batch suspension culture. From these two categories, clonal ranking was highlighted within all phase
platforms including (i) spot test, (ii) 24-well plate, (iii) batch culture and (iv) Fed-batch culture. This
figure was sourced from (Porter et al. 2010).
In the same study by Porter and colleagues (Porter et al. 2010), the addition of a feed
resulted in the largest variation (25%) in ranking position across all clones. This was
observed in the case of all miR-23 sponge and miR-NC sponge clones cultured with a
feed supplement causing larger variations in growth and maximum cell density (Fig. 3.58
A). Volumetric SEAP production was still retained at 3-fold increase when compared to
non-specific controls (Fig. 3.58 A). However, when volumetric SEAP activity was
296
compared to that of miR-23 sponge clones cultured under a 6 day batch, it was not
improved, in fact mildly reduced, (Fig. 3.56 C and 3.58 C) despite the significant increase
in maximal cell density. One report observed that in the case of CHO cells producing
tissue plasminogen activator (t-PA), the reduced volumetric yield associated with a 2.5-
fold increase in IVCC was mediated by a possible secretory bottleneck when intracellular
t-PA mRNA levels were identified to be increased (Dowd, Kwok & Piret 2000). Another
possibility is that the enhanced cellular growth has resulted in a trade off in CHO cell
specific productivity conferring no significant production advantage despite the extension
of 2 days in the culture life time. Both miR-23 sponge clones subject to a nutrient feed
also retained a viability above 80% for an additional 48 h compared to control clones.
The introduction of nutrient feeds have greatly improved pharmaceutical recombinant
protein manufacturing by boosting maximal cell densities, improving cell productivities
and enhancing viability (Butler, Meneses-Acosta 2012, Huang et al. 2010). This enhanced
cell viability was evident in both miR-23 and miR-NC sponge clones upon fed-batch
cultivation when compared to batch, however, stable miR-23 depletion did appear to
confer a further extended viability (Fig. 3.58 B).
Low culture temperature results in growth arrest, meaning that cell density is significantly
reduced at 31°C, nutrients are consumed at a slower rate and the culture can be run for a
longer period of time prior to the onset of cell death, in a cell-type specific manner. We
thought it prudent to exploit the potential enhanced viability phenotype of a fed-batch
with that of a temperature shift as a means to boost volumetric SEAP production. When
specific productivity was compared between miR-23 sponge clones 4 and 6 cultured
under fed-batch conditions with or without temperature shift to 33°C (Fig. 3.60), there
was an average ~2-fold increase in cell specific SEAP productivity in the case of
hypothermic culture conditions. Furthermore, volumetric productivity was enhanced by
~30% in the case of clones subject to temperature shift with feed addition (Fig. 3.58 C
and 3.59 C). A similar observation was made by Fox and colleagues in the case of
Interferon-γ production (Fox et al. 2004). The highest volumetric SEAP production
observed in the case of fed-batch with temperature shift may be due to the extended
production phase. Another possibility that has been reported is that reduced culture
temperatures can have the added advantage of stabilizing the product and reducing
intramolecular aggregation (Tharmalingam, Sunley & Butler 2008, Oguchi et al. 2006).
The highest level of SEAP activity was recorded under hypothermic growth conditions.
Although, the presence of the miR-23 sponge in fed-batch appeared to delay the onset of
297
apoptosis, the extended production phase under hypothermic growth conditions appeared
to be maintained below the industry standard of 80%, with the anti-apoptosis benefits
appearing from day 11 onwards (Fig. 3.59 B).
It is possible that clones under batch conditions will exhibit a unique transcriptome to a
clone grown in fed-batch culture. This could ultimately provide a completely new set of
targets for miR-23 impeding its role in regulating particular phenotypes. It was found that
cell death in fed-batch was primarily induced via death receptor- and mitochondria-
mediated signalling pathways rather than endoplasmic reticulum-mediated signalling
compared to batch (Wong et al. 2006) with Fas being demonstrated to be a target of miR-
23a/b (Li et al. 2013a). Despite miR-7 over-expression yielding no negative impact on
CHO cell viability (Barron et al. 2011b) and being associated with the de-regulation of
pro/anti-apoptotic genes (Sanchez et al. 2013a), its stable inhibition mediated a prolonged
viable phenotype under fed-batch conditions (Sanchez et al. 2013b) again suggesting the
availability of a new cohort of targets. It is possible that fed-batch with temperature shift
would best suit the stable depletion of miR-23 in our CHO-SEAP population with further
optimisation such as adapting miR-23 sponge clones to hypothermic growth conditions
(Sunley, Tharmalingam & Butler 2008b).
4.1.6 miR-23 depletion enhances CHO cell mitochondrial function at Complex I and
II
Mitochondria are vital to normal cellular function as they are responsible for energy
production in eukaryotes, including the synthesis of phospholipids and heme, calcium
homeostasis and apoptosis activation. Deregulation in mitochondrial activity is often
associated with numerous disease states such as diabetes mellitus. Furthermore,
alterations in the metabolic pathways that ultimately feed oxidative phosphorylation
(OXPHOS) at the electron transport system such as glycolysis and the tricarboxylic acid
cycle (TCA) can lead to changes in cell fate. The Warburg effect, a process also observed
in CHO cells (Martinez et al. 2013), has been observed to be reprogrammed through miR-
23’s interaction with glutaminase and sparked the exploration of mitochondrial activity
(Gao et al. 2009). Each of the enzymes that constitute the glycolytic pathways have been
found to be over-expressed and/or deregulated in several cancer cell lines and tumours in
vivo (Pelicano et al. 2006a) ultimately contributing t(Martinez et al. 2013)o enhanced cell
growth. However, in this instance, miR-23 depletion resulted in enhanced SEAP
298
transcription without contributing to growth. A recent study demonstrated the metabolic
flux in an antibody producing CHO cell line exposing variations in metabolism during
growth (Glycolytic) and production (TCA) (Templeton et al. 2013). This has further been
supported with observations that increased oxidative metabolism can be favoured in
situations of nutrient depletion and/or reduced growth rate (Young 2013). These studies
implicate certain metabolic states with particular CHO cell phenotypes, mainly
productivity with oxidative metabolism encouraging the exploration of miR-23 depletion.
Eliminating proliferation as a variable, it has been observed that the protein folding
process consumes much energy, which is provided for by the mitochondria (Bravo et al.
2011, Simmen et al. 2010). With this in mind, enhanced mitochondrial function could
further aid miR-23 depleted clones in efficiently processing and turning over the
additional SEAP mRNA transcripts (Fig. 4.1).
When intact, non-permeabilised miR-23 clones were assessed for their mitochondrial
function, one set of clonal pairs exhibited an increased Routine and Leak respiration due
to the reduced calculated maximum ETS when applying the respiratory control ratios
(RCRs) (Fig. 3.71 B). RCRs are defined as the O2 flux in different respiratory control
states normalised for the maximum flux of the ETS capacity. Despite an overall reduced
maximum ETS in the case of miR-23 sponge clone 4 compared to miR-NC sponge clone
2, energy turnover is potentially more efficient. Muscle cell mitochondria have
demonstrated increased RCRs indicating efficient energy production (Larsen et al. 2011).
The proton gradient is the source of electrochemical energy and regulates electron
movement across the transport system, electrochemical backpressure (Fig. 4.4).
299
Figure 4.4: Schematic diagram of the mitochondrial proton gradient
Figure 4.4: The propagation of electrons through the ETS is regulated by the electrochemical
backpressure exerted by the proton gradient. Protons are translocated from the mitochondrial
matrix to the inner mitochondrial membrane through the various complexes (I, III and IV) as
electrons are donated from NADH/FADH2. Inner membrane protons are then translocated back into
the matrix through ATP Synthase generating ATP or lost to proton leak.
http://hyperphysics.phy-astr.gsu.edu/hbase/biology/etrans.html.
However, both protons and electrons can leak across the membrane without contributing
to ATP production yet ultimately allowing electron transport from the reducing
equivalents (NADH and FADH2). Leak respiration is dependent on internal energy
sources as ATP Synthase has been inhibited by the addition of oligomycin thus preventing
the translocation of exogenous metabolites into the mitochondrial matrix. Leak
respiration also accounts for a large proportion of resting metabolic rates (20-25%) (Rolfe
et al. 1999). The observed increase in leak respiration indicated that internal substrates
for mitochondrial turnover are more readily available. This is in contrast to the other
clonal pair, miR-23 sponge clone 6 and miR-NC sponge clone 11, whose RCRs remained
similar while miR-23 sponge clone 6 demonstrated an enhanced routine respiration and
ETS capacity (Fig. 3.72). This would suggest that mitochondrial function may not be
contributing directly to an increase in the transcriptional activity of the SEAP mRNA
transcript. However, an increase in cellular ATP levels can activate transcription from
promoters by RNA polymerase II (Conaway, Conaway 1988) indicating a role in
transcriptional control. To put the change of 5% difference in Leak respiration into
perspective, a study reported a difference of 5% in leak respiration in skeletal muscle
300
mitochondria in diet resistant and diet responsive patients (Harper et al. 2002), i.e. 5%
would be considered significant as it does not vary hugely.
Systematic stimulation and inhibition of the various complex components of the electron
transport system, when cells were permeabilised with Digitonin, revealed that there was
a 30% increase in oxygen consumption attributed to Complex I and II (Fig. 3.74 B). This
indicated that there was an increase in the availability of the reducing equivalents NADH
and FADH2 which provides electrons that drive the pumping of protons from the matrix
to the inner membrane space contributing to energy production or a more effective and
efficient assembly of these components on the inner mitochondrial membrane (Fig. 4.5).
Cell-to-cell variation in the rate of transcription has been correlated to higher
mitochondrial mass or higher membrane potential, with the presence of anti-/pro-oxidants
and increased ATP substantially improving transcription rates (das Neves et al. 2010),
indicating the possibility of a contribution to enhanced SEAP transcription. However, this
enhanced transcription was reported to be on a global level which we investigated by
probing a set of genes (GLS, XIAP and TFAM). These showed no increase in miR-23
sponge clones. However, in hindsight, if global gene transcription was enhanced then so
would the transcription of the endogenous control used for comparison in qPCR
indicating that no difference would be detected.
A kinetic model of cellular energetics in CHO cells subjected to sodium butyrate exposure
demonstrated a shift in energy metabolism towards increased efficiency of glucose
utilisation through the TCA cycle (Ghorbaniaghdam, Henry & Jolicoeur 2013). As
sodium butyrate treatment is routinely used to artificially induce higher specific
productivity, the 30% increase in mitochondrial function observed in the miR-23 sponge
clones supports this notion of oxidative metabolism being linked to specific productivity.
NaBu treatment has also been observed to stimulate the release of Ca2+ from the ER (Sun
et al. 2012a) with Ca2+ itself being previously demonstrated to activate enzymes of the
TCA cycle thus enhancing OXPHOS (Osellame, Blacker & Duchen 2012). It still remains
difficult to speculate as to how there appears to be a selective increase in SEAP
transcription. This increase of 30% is significant as reports in the literature of variations
in mitochondrial function are mild, usually in the range of 5-70% (Hirsch et al. 2012).
301
Figure 4.5: Schematic Diagram linking the TCA cycle to ATP production
Figure 4.5: The schematic representation of the complex components of the electron transport system
(ETS) in addition to its coupled behaviour with the TCA cycle and the reducing equivalents produced
(Osellame, Blacker & Duchen 2012).
Despite the potential increase in mitochondrial function, there did not seem to be an
apparent compromise in mitochondrial membrane integrity. Complex I and III of the
electron transport system have been described as the primary location of electron loss or
302
leak (Jastroch et al. 2010). This could potentially contribute to the lack of synergy
observed upon co-stimulation of both complexes I and II when ADP was not the limiting
factor. Both CI and CII demonstrated a ~30% increase in oxygen consumption
independently, yet maintained this increase upon co-stimulation. This could potentially
be attributed to electron leak downstream of both complexes. Mitochondrial membrane
integrity was demonstrated to be unaffected in miR-23 depleted clones both through the
addition of cytochrome c during HRR-SUIT permeabilization and the evaluation of
mitochondrial content/mass using the MitotrackerTM deep red assay. MitotrackerTM deep
red probes are sensitive to membrane integrity and will diffuse out which we assessed not
to be a contributing factor.
Upon inhibition of the activity of complex II and III by Malonic acid and Antimycin A,
respectively, there was an observed drop in oxygen consumption on complex IV for miR-
23 depleted clone upon its stimulation by the artificial electron donor TMPD. This again
could potentially be another contributing factor to the lack of synergy between the
stimulation of complex I and II with electrons potentially reaching a bottle neck at
complex IV.
When we assessed the levels of glutamate, there was an observed increase of 65% across
all clones (Fig. 3.70). Assessment of this metabolite was suggested by the identification
of the role that miR-23a/b plays in targeting the enzyme glutaminase (Gao et al. 2009).
The process of glutaminolysis is an anapleurotic reaction that provides that TCA cycle
with vital intermediates such as α-ketoglutarate (Fig. 4.5). This is a potential mechanism
in which cancer cells can overcome what’s been termed “the Crabtree effect” in which
excessive turnover of glucose via the “Warburg effect” can quench and ultimately starve
the TCA cycle of Acetyl-CoA through exclusive conversion of pyruvate to lactate
(Dell'Antone 2012). This increase in glutamate availability could explain the increase in
respiration observed at complex I and II during permeabilised assays as well as the
increase in routine and leak respiration with the potential of an alternative endogenous
source of TCA cycle substrates.
Peak antibody production in CHO cells has been associated with a highly metabolic
oxidative state while peak cell growth was characterised by a highly glycolytic state
(Templeton et al. 2013). This would suggest that the observed SEAP productivity could
be attributed to enhanced mitochondrial function also indicating why cellular
proliferation was not influenced. However, it would be necessary to assess the glycolytic
303
status of miR-23 sponge clones to rule this a plausible explanation. Glutamate
dehydrogenase (GDH) can also convert α-ketoglutarate into glutamate providing an
alternative source of glutamate that is not derived from glutaminolysis (Friday et al.
2012). Interestingly, when proteomic analysis was carried out between miR-NC and miR-
23 sponge clones, glutaminase was not identified to be DE suggesting that the formation
of glutamate could be as a result of catapleurosis derived from downstream sources to
TCA intermediates.
These results suggest that miR-23 depletion is potentially playing a role in enhancing
CHO cell mitochondrial function. The link between oxidative metabolism and
productivity is a strong indication that the improved mitochondrial function is influencing
SEAP transcription. Without assessing ATP levels it would be premature to equate this
increased oxygen consumption to energy synthesis. Furthermore, identifying and
comparing the presence of lactate would indicate if CHO cells are favouring the glycolytic
pathway which could be an explanation as to why miR-23 sponge cells demonstrate no
growth advantage.
4.1.7 Summary of miR-23 depletion in CHO cells
miR-23 depletion appeared to influence the CHO cell phenotype in a product-dependent
manner as both SEAP and IgG secreting cell lines were generated from the same CHO-
K1 parent. miR-23 depletion in CHO-1.14 cells improved CHO cell growth and extended
cell viability at the expense of IgG productivity. This phenotype was maintained as far as
clonal isolation at which point particular control clones exhibited a comparable
phenotype. Depletion of miR-23 in CHO-SEAP cells improved SEAP productivity to a
maximum average of 3-fold without influencing growth. Under fed-batch culture
conditions, miR-23 depletion was observed to prolong the onset of cell death. Improved
specific SEAP productivity was shown to be, in part at least, at the transcriptional level
with cell size and mRNA stability being ruled out as contributing factors. miR-23 sponge
clones exhibited an enhanced mitochondrial function with permeabilised assays revealing
a hyperactivity of 30% at Complex I and Complex II. The difference in mitochondrial
mass/content between miR-23 and miR-NC clones was demonstrated to be negligible.
However, the specificity of miR-23 depletion influencing SEAP transcription remained
undetermined at this point despite the indication of oxidative metabolism being associated
with specific productivity.
304
4.2 Biotinylated miRNA mimics as a novel method for mRNA target identification
A novel method of isolating and identifying potential miRNA targets was carried out by
exploiting the use of biotin-labelled miRNAs. Previous studies have demonstrated the
utility of selective enrichment of target mRNAs associated with a specific miRNA such
as the immunoprecipitation of epitope-tagged Ago2 in HEK293 cells transfected with
miR-124a (Karginov et al. 2007). The inherent benefits of using biotin-tagged miRNA
would be the identification of direct targets, excluding genes whose expression is
indirectly influenced by changes in miRNA expression, a refinement not possible in
microarray or proteomics of whole cell lysates. Furthermore, affinity pull-down would
isolate mRNAs that exhibit a binding affinity for a selected miRNA. Upon assessing the
utility of this technique using miR-23b*, we carried out pull-down followed by
microarray target identification, using CHO-specific arrays, for both biotin-tagged miR-
23b and miR-23b*. When both datasets were compared to that of the biotin-tagged non-
specific control, no enrichment was detected (no hierarchal clustering of samples) when
data were assessed for trends.
miR:miR* partners have been demonstrated to have complementary functionality be it
through the same mRNA targets or similar signalling networks (Yang et al. 2013); or
unique functionality through an independent set of mRNA targets (Shan et al. 2013). By
comparing the targets pulled down with miR-23 and miR-23b*, we sought to shed light
on the relationship and potential functional redundancy between both miRNAs.
Furthermore, we designed a biotin-labelled miR-23b* (miR-23b*-2 modified) whose
duplex gave rise to both miR-23b and miR-23b* during processing. The biotin-linker was
located on the miR-23b* passenger strand and was intended to examine the fate of a
“discarded” passenger strand when compared to a biotin-linked miR-23b* favoured for
selection and RISC-loading. However, only 12 CHO-specific array cards were available
and the inclusion of both cells only and biotin-labelled non-specific control was essential
resulting in the utilisation of both miR-23b and miR-23b* biotin-labelled mimics.
As previously mentioned, miR-23 possesses an inherently weak seed-pairing due to its
AT-rich seed region (Garcia et al. 2011). Despite the gentle lysis process, carried out to
avoid disturbing any ribonucleoprotein complexes that would otherwise fall apart in
harsher protocols, this weak pairing could have contributed to the poor pull-down of miR-
23-specific mRNAs resulting in the isolation predominantly of non-specific mRNA
transcripts. When qRT-PCR was carried out, it was observed that there was no enrichment
305
detected for miR-23b in either streptavidin-bead-incubated lysates as well as input
samples prior to streptavidin incubation. It was possible that the biotin-tagged miR-23b
mimic was not being processed efficiently. Keeping in mind the potential for poor
transfection efficiency on the day, with a 14-fold enrichment of miR-23b*, it was possible
that this could have further contributed to the absence of enrichment of miR-23b
ultimately resulting in no capture of target mRNAs.
The over-expression of biotin-tagged miR-23b*-1 did not mediate the phenotype
previously observed for the transient expression of miR-23b* during the primary miRNA
screen reported in section 3.1 of results (reduced cell growth and viability). This
observation would suggest that the biotin-labelled miR-23b*-1 was not being sufficiently
processed in the cell to yield the mature biotin-labelled, RISC-loaded, miR-23b* that
would otherwise interact with endogenous miR-23b*-specific transcripts mediating their
degradation or translational suppression. Despite this potential lack of cellular processing
of the miR-23b*-1 miRNA, enrichment and quantification was observed both in the
optimisation process and the main experimental run. It is possible that regardless of
efficient processing, the RNA extraction process would separate the exogenously
introduced miRNA duplex leaving the miR-23b* strand available for qRT-PCR
amplification and detection. This would explain how a high fold-change of ~600 was
identified for miR-23b* during optimisation. However, a ~14-fold change was detected
for samples that were subsequently loaded on to microarrays. It is possible that the
reduced enrichment of miR-23b* compared to the enrichment achieved during
optimisation was due to a poor transfection efficiency on the day ultimately contributing
to no mRNA target enrichment when microarrays were hybridised and analysed. Lal and
colleagues reported a >40-fold enrichment of miR-34a in a similar experiment where 982
genes were found to be captured (ranging from 4-10-fold) and common to two cell lines
including the identification of 14 new targets of miR-34a (Lal et al. 2011). One of the
genes predicted to be a target of miR-23b* was phosphoglycerokinase (Pgk-1). It was the
detection of this gene that gave us the most confident indication that we were achieving
selective enrichment with miR-23b* (Dheda et al. 2004). On the other hand, its role as a
housekeeping gene and therefore relatively high and consistent expression could have
contributed to its identification during optimization, when it was potentially a
contaminant or non-specifically bound.
306
At the time, the lack of characterisation of miR-23b* and the limited knowledge of bona
fide targets impeded the validation and optimisation of this method of identifying targets
of miR-23b and miR-23b*. Previous studies implementing this technology had the
fundamental resource of validated targets in the cell lines under investigation such as the
miRNA bantam and its gene target Hid in Drosophila (Orom, Lund 2007) and CDK4/6
for miR-34a in HEK293 cells (Lal et al. 2011). In future, optimisation of this process
would be essential with a target previously validated in CHO cells or validated prior to
using this biotin capture approach. Furthermore, if numerous biotin-tagged miRNA were
being explored, it would be essential to validate each miRNA independently for criteria
such as efficient cellular processing, execution of an expected phenotype following its
over-expression, qPCR enrichment for mature miRNA levels and finally mRNA target
association and enrichment. Although this method was unsuccessful in our case, we
propose several viable reasons as to the potential reasons behind its failure. Nonetheless,
it still remains an attractive approach for the identification of miRNA specific mRNA
interactions.
4.3 Analysis of the whole cell proteome reveals potential miR-23 regulated processes
As miRNA regulation ultimately culminates at the interference of mature protein levels
due to post-transcriptional suppression/de-repression, label-free mass spectrometry was
carried out as a means to identify finite changes in the translatome of miR-23 depleted
CHO-SEAP clones demonstrating enhanced specific productivity and mitochondrial
function. 41 proteins were identified to be DE in miR-23 depleted clones when compared
to control CHO clones. The influence of miR-23 diversion is of importance when
attempting to identify potential candidates under miR-23s influence in addition to which
of these candidates are contributing to the observed phenotype. For this reason, a small
sub-list of DE proteins are discussed below that could potentially be involved in the
observed elevated SEAP transcription in addition to enhanced mitochondrial activity.
4.3.1 LETM1
LETM1 (Leucine zipper-EF-hand-containing transmembrane protein) is a nuclear
encoded mitochondrial inner membrane protein that is involved in respiratory chain
biogenesis and calcium (Ca2+) homeostasis in mitochondria. It has been implicated in the
307
onset of Wolf-Hirschhorn syndrome in which its deletion mediates an impaired
mitochondrial ATP generation, reduced glucose oxidation, altered metabolism and
increased seizures (Jiang et al. 2013a). We identified LETM1 to be up-regulated by 2.73-
fold in CHO-SEAP cells engineered with a sponge specific against miR-23.
Oxidative phosphorylation is regulated by cross-talk with cellular calcium signalling, so
that the transfer of Ca2+ into the mitochondrial matrix signals an increased energy demand
(Jacobson, Duchen 2004). The rise in matrix Ca2+ concentration [Ca2+]m activates the rate
limiting enzymes of the TCA cycle – pyruvate dehydrogenase, α-ketoglutarate
dehydrogenase and NAD-isocitrate dehydrogenase as well as ATP synthase although the
mechanism remains unclear (Osellame, Blacker & Duchen 2012). These processes
ultimately drive the increase in the availability of NADH to the respiratory chain (a
substrate for complex I), an increase in respiration and ATP synthesis. LETM1 has been
characterised as a Ca2+/H+ antiporter contributing to Ca2+ influx into the mitochondria
along with other Ca2+ regulatory genes such as the mitochondrial calcium uniporter
(MCU) (Fig. 4.6).
308
Figure 4.6: LETM1s function as a Ca2+/H+ antiporter in the mitochondria
Figure 4.6: Illustration of the various methods exploited by the mitochondria to import cytosolic Ca2+
into the mitochondrial matrix with LETM1 acting as a Ca2+/H+ antiporter. This figure was sourced
from (Pan, Ryu & Sheu 2011).
With the previous observation that Ca2+ drives mitochondrial energy demand, it is
possible that LETM1 overexpression contributes to the enhanced OXPHOS observed.
Furthermore, RNAi-mediated inhibition of LETM1 was demonstrated to nearly abolish
both mitochondrial Ca2+ and pH increases which indicated that LETM1 catalysed the slow
uptake of Ca2+ into the mitochondria (Jiang, Zhao & Clapham 2009). The study also
demonstrated a sustained Ca2+ increase upon LETM1 inhibition suggesting that Ca2+
efflux is highly regulated by this protein. Additionally, as the formation of the proton
gradient by Complex I, III and IV is vital for harnessing electrochemical energy for the
synthesis of ATP, it has further been suggested that the uptake of Ca2+ into the matrix by
LETM1 not only drives energy demand by the activation of TCA cycle enzymes but also
contributes to the electrochemical gradient ending at ATP production via the export of H+
out of the matrix (Poburko, Demaurex 2012) (Fig. 4.1).
309
Another interesting role of LETM1/mdm38 (its Saccharomyces cerevisiae homolog) is its
capability to bind mitochondrial ribosomes through a conserved 14-3-3-like ribosomal
binding domain (RBD) (Lupo et al. 2011). Enhanced respiratory chain biogenesis has
been observed to be specific to the effective translation and synthesis of the complex III
component cytochrome b reductase (Cyt-b) and the complex IV component cytochrome
c oxidase (COX1) (Bauerschmitt et al. 2010) mediated by LETM1/mdm38 while its loss
leads to respiratory chain defects. Interestingly, when we assessed mitochondrial function
in permeabilised CHO cells, a ~30% increase in oxygen consumption was observed for
complex I and II while a drop in oxygen consumption was observed at complex IV in the
case of the miR-23 sponge clone 4. It has been demonstrated that overexpression of
LETM1 can abrogate mitochondrial chain biogenesis (Piao et al. 2009) suggesting a
potential reasoning behind this reduced activity at complex IV and the lack of synergy
between complex I and II. LETM1 overexpression has additionally been documented in
various cancerous phenotypes and its inhibition of mitochondrial function has correlated
with an increase in glycolysis, compensating for ATP production and resulting in a shift
from NADH to NADPH production (Chen et al. 2007, Pelicano et al. 2006b). Although
components specific to the mitochondrial transport chain were not found to be DE in miR-
23 depleted clones, it is possible that their assembly is being perturbed. Furthermore,
mitochondrial enrichment and label-free mass spectrometry was carried out in an attempt
to identify subtle alterations taking place within the mitochondria that may have been
missed through whole cell proteomic analysis. The enhanced mitochondrial function may
be as a result of the activation of particular TCA cycle enzymes as suggested by LETM1s
role in Ca2+ homeostasis and the observed up-regulation of metabolic enzymes such as
isocitrate dehydrogenase (IDH1) and malate dehydrogenase (MDH) (Fig. 4.1). With
reference to the literature, LETM1 appears to have a regulatory influence in mitochondrial
function both directly and indirectly spanning the majority of the respiratory chain
complex components.
4.3.2 14-3-3-eta (YWHAH)
The 14-3-3 protein family is a well conserved family of heterogeneous dimeric proteins
whose activation is dependent upon calcium and the cAMP-dependent kinase or
Calmodulin-dependent protein kinase. They are involved in cell signalling, regulation of
cell cycle progression, intracellular trafficking/targeting, cytoskeletal structure and
310
transcription and commonly require the phosphorylation of a serine or threonine residue
in their target sequences (Ferl, Manak & Reyes 2002). There are 7 known isoforms in
mammalians (α/β, ε, η, γ, τ/θ, δ/ζ, and σ) suggesting that each exert a distinct function.
However, knockout studies have demonstrated no definite phenotype which may reflect
the fact that in many cases isoforms can compensate for each other (Aitken 2006).
14-3-3-eta (14-3-3-η) was found to be elevated in its abundance in miR-23 depleted CHO-
SEAP cells by 2.7 fold. The regular functions of 14-3-3 proteins encompasses activation,
inhibition, stabilization or orientation of their binding partners (Gardino, Smerdon &
Yaffe 2006). This is through the sequestration of target proteins in the cytoplasm
ultimately inhibiting nuclear or mitochondrial localisation of specific binding partners
and includes cellular functions such as metabolism (Kleppe et al. 2011) and insulin
secretion (Chen et al. 2011b).
Although we are not observing a tumour-like phenotype as previously demonstrated for
this family of proteins, one interesting role is their involvement in the regulation of gene
expression, reviewed in (Healy, Khan & Davie 2011). 14-3-3-epsilon’s expression in
CHO has been previously demonstrated to be induced upon the induction of hypothermic
growth conditions (Thaisuchat et al. 2011) potentially implicating a role in increased
specific productivity (Kumar, Gammell & Clynes 2007). Additionally, 14-3-3-epsilon,
among other 14-3-3 family members, was identified to be increased in their
expression/translation upon transient transfection with pre-miR-7 (Meleady et al. 2012a)
whose transient overexpression was also found to mimic a temperature-shift-like
phenotype including enhanced specific productivity (Barron et al. 2011b). This would
suggest that the expression of this family of 14-3-3 proteins is involved in enhancing
transcription in a temperature-dependent and independent manner.
It has been further demonstrated that the 14-3-3 family of proteins can interfere with gene
expression at an epigenetic level by binding to and sequestering (Shahbazian, Grunstein
2007)HDAC4 activity (Wang et al. 2000) and HDAC5 (Grozinger, Schreiber 2000).
Histone deacetyltranferases (HDACs) regulate gene expression by the phosphorylation of
H3 of the histone complex rendering surrounding DNA inaccessible. Although the
association with HDAC4 and 14-3-3 does not interfere with HDAC4s activity, retention
within the cytoplasm prevents its nuclear localisation and ultimately inhibition of
transcriptional repression mediated through deacetylation (Fig. 4.1). Additionally,
prevention of nuclear entry of HDAC4 via 14-3-3 was demonstrated to be activated upon
311
Ca2+/Calmodulin signalling, a biological process found to be activated within miR-23
depleted CHO clones (Guan et al. 2012) including CaMKIIδ found also to be increased
in expression in miR-23 depleted CHO clones (Zhang et al. 2007). HDACs lack intrinsic
DNA-binding activity and are therefore recruited to target sites by their direct association
with transcriptional activators and repressors (Shahbazian, Grunstein 2007). Target site
specificity to the SEAP CMV-promoter and enhanced SEAP transcription in the case of
miR-23 sponge clones is partially suggested when considering the binding specificity can
be dictated by endogenous transcriptional activators that would be cell type specific yet
potentially similar within our rCHO-SEAP cell line. To explore the possibility of 14-3-3
mediating specific transcriptional enhancement of the SEAP gene further validation
would be required such as siRNA-mediated knockdown of the 14-3-3-eta gene expecting
to observe reduction in SEAP secretion and transcription or immunoprecipitation of 14-
3-3-eta as means to identify its binding partners, HDAC4 potentially being one.
Interestingly, a recent paper by Yang and colleagues demonstrated that the addition of the
small molecule inhibitor of HDACs, Valporic acid (VPA), enhanced antibody titres by
>20% in numerous cell lines without impacting on glyco-profiles (Yang et al. 2014).
Several miRNA have been previously demonstrated to regulate the translation of 14-3-3
proteins in various types of cancer such as miR-193b in MCF7 cells (Leivonen et al. 2011)
and miR-375 in gastric cancer (Tsukamoto et al. 2010). Interestingly, miR-375 has been
widely implicated in the regulation of glucose-stimulated insulin secretion through
targeting multiple factors such as PDK1 and myotrophin in pancreatic β-cells (Li 2014).
Its previously characterised impact on 14-3-3 proteins expression offers a potential and
alternative explanation to how it mediates elevated levels of insulin transcript. Despite
the large regulatory capabilities of this family of proteins, it is possible that 14-3-3 is
mediating the enhanced SEAP transcription phenotype be it through direct de-repression
due to miR-23 diversion or as a result of other targets more dramatically de-repressed
such as CaMKIIδ.
4.3.3 Calcium signalling
This ubiquitous second messenger, Calcium (Ca2+), is intricately involved in numerous
cellular processes including signal transduction, muscle contraction, secretion of
proteins/hormones and gene expression. Within the cell, Ca2+ is stored in specialised
compartments (endosomes, lysosomes and golgi) as well as within the endoplasmic
312
reticulum (ER)/sarcoplasmic reticulum (SR) (Rizzuto, Pozzan 2006). Protein folding in
the ER is a very energy-demanding process as many of the molecular chaperones (BiP
and GP94) hydrolyse ATP during their binding/release cycles. It has been suggested that
the depletion of intraluminal ER ATP may be an energy deprivation signal to stimulate
Ca2+ release for uptake by the mitochondria ultimately stimulating OXPHOS as described
earlier through the activation of TCA cycle enzymes (Kaufman, Malhotra 2014). This
would suggest that enhanced transcription of the SEAP gene within our miR-23 depleted
clones is exerting a high energy demand on the ER for efficient folding and processing
ultimately driving energy demand through Ca2+ release and mitochondrial stimulation
(Fig. 4.1). Excessive cytosolic/mitochondrial Ca2+ can result in mitochondrial swelling,
pore opening and collapse of the mitochondrial membrane potential resulting in apoptosis
by cyt-c release (Osellame, Blacker & Duchen 2012). This would suggest that if
mitochondrial function is being stimulated by Ca2+ influx, it is as a result of energy
demand and not the direct influence of miR-23 intervention providing unphysiological
amounts of Ca2+ through the over-activity of LETM1, for example. An additional layer or
Ca2+-mediated OXPHOS regulation comes from the observation that extra-mitochondrial
Ca2+ can influence the transport of glutamate into the mitochondria (Gellerich et al. 2010),
suggesting the observed increased availability of glutamate is being channelled into the
mitochondria for utilisation.
CaMKs (Ca2+/Calmodulin-dependant protein kinases) are a family of serine/threonine
kinases regulated by calcium and Calmodulin. One member, CaMKIIδ, was observed to
be increased in expression by 1.75-fold in miR-23 depleted clones. Another protein found
to be up-regulated and activated by CaMKs was Caldesmon (CALD), an actin binding
protein (2.16-fold). Both CaMKIIδ and CALD presented themselves as potential targets
of miR-23 and both are responsive to calcium signalling.
CaMKIIδ has previously been found to be expressed at relatively high levels in β-cells
(Easom 1999) in addition to its activation upon glucose stimulation following the entry
of Ca2+ (Norling et al. 1994, Wenham, Landt & Easom 1994). Furthermore, CaMKIIδ
was found to be directly related to the metabolic control of insulin secretion at a
transcriptional level (Osterhoff et al. 2003). As mentioned previously, CaMKIIδ has been
demonstrated to phosphorylate the two conserved phosphorylation sites, ser-246/467 in
HDAC4, ultimately flagging it for the mobilisation of 14-3-3-eta which recognises
phosphorylated serine/threonine residues (Little et al. 2007). Retention of HDAC4 in the
313
cytosol prevents histone deacetylation maintaining DNA in a relaxed and
transcriptionally accessible state (Fig. 4.1) Furthermore, CaMKIIδ possesses two splice
variants (δB and δC) the former of which contains a nuclear localisation signal (NLS)
(Zhang, Brown 2004) further suggesting its potential role in enhancing SEAP
transcription. Perhaps elevated SEAP productivity could be as a result of this process with
mitochondrial function being a by-product of calcium signalling stimulated by ATP-
depleted ER-release. It would be interesting to evaluate the levels of intracellular calcium
in our miR-23 stable clones in addition to inhibiting CaMKIIδ.
4.3.4 TCA cycle and metabolism
Several proteins were found to be DE in CHO-SEAP-sponge-23 that could potentially be
playing a role in cellular metabolism including isocitrate dehydrogenase 1 (IDH1-1.68-
fold up regulated), malate dehydrogenase (MDH – 1.67-fold up regulated), Lysosomal
alpha glucosidase (1.65-fold up regulated) and prolyl-4-hydroylase (P4HA1 – 1.84-fold
up-regulated). It’s been suggested that some of these candidates can be activated by an
influx of Ca2+ into the mitochondrial matrix suggesting their function to be responsive to
internal cellular cues and potentially not causative. Nevertheless, their increased
abundance could be responsible for the increase in mitochondrial OXPHOS previously
observed in miR-23 depleted CHO-SEAP cells (Fig. 4.1).
Lysosomal alpha glycosidase is an enzyme encoded by the GAA gene that mediates the
degradation of glycogen to glucose from lysosomes. Lysosomal alpha-glucosidase is
known to be responsible for a lysosomal storage disease, Pompe(Manganelli, Ruggiero
2013) disease, whereby a build-up of lysosomal glycogen results in the inhibition of
lysosome activity, aberration of autophagy and ultimately muscle wasting characterised
by the build-up of cellular debris (Manganelli, Ruggiero 2013). Mitochondrial
dysfunction has also been implicated in the onset of Pompe disease, however, this lull in
mitochondria function is potentially coupled to the inhibited autophagic turnover of
cellular organelles in addition to broken cristae due to the accumulation of membrane
bound glycogen particles within the mitochondria (Raben et al. 2012). Increased
expression of this enzyme in miR-23 depleted CHO cell could potentially supply excess
glucose through glycolysis into the TCA cycle resulting in the observed increase in
OXPHOS in addition to ensuring effective and optimal lysosome function.
314
IDH is a family of enzymes that include three distinct enzymes: IDH1, IDH2 and IDH3.
All three enzymes catalyse the conversion of isocitrate to α-KG, a central metabolite for
the TCA cycle. IDH1 and 2 use NADP+ while IDH3 uses NAD+ with IDH1 being located
in the cytosol and the latter two being localised to the mitochondria (Yang et al. 2012).
As NADH is a reducing equivalent providing electrons to the ETS and IDH3s reaction
being irreversible, this isoform is considered central to oxidative phosphorylation (Fig.
4.7).
Figure 4.7: The role of IDH in cellular metabolism
Figure 4.7: The role of the IDH family of enzymes in cellular metabolism including the central role
IDH3 plays in the TCA cycle and both the compensatory roles IDH1/2 play in the TCA cycle in
addition to downstream molecular pathways. This figure was sourced from (Yang et al. 2012).
NADPH on the other hand is involved in many cellular processes including defence
against oxidative stress, fatty acid synthesis and cholesterol biosynthesis (Kim et al.
2007). It is possible that the cytosolic IDH1 enzyme reaction might compensate for the
loss of the mitochondrial IDH2/3 by producing metabolites in the cytosol that can be
transported into the mitochondria as it has previously been observed that rapid transport
of metabolites can take place between the cytosol and the mitochondria such as citrate,
isocitrate and α-ketoglutarate (Fig. 4.8).
315
Figure 4.8: The role of IDH1 in supplementing the TCA cycle
Figure 4.8: The potential role of IDH1 in rCHO-SEAP cells inducing an increased OXPHOS. IDH1
catalysis the reversible conversion of isocitrate to α-KG with the forward reaction being
predominant. This reaction takes place in the cytoplasm whereby α-KG is then transported into the
mitochondria for entrance into the TCA cycle in what’s known as anaplerosis.
Mitochondrial export of citrate/isocitrate was demonstrated to be an important step in
glucose-stimulated insulin secretion with the overexpression of citrate/isocitrate carriers
(CICs) enhancing GSIS (Joseph et al. 2006). Furthermore, siRNA knockdown of
cytosolic IDH1 inhibited GSIS by 59% (Ronnebaum et al. 2006). This observation
reinforces the suggested function of α-KG as a co-factor for other cellular pathways such
as the activation of α-KG-dependent dioxygenases and that the predominant mutations in
IDH1/2 isotypes influence these alternative molecular pathways. Although IDH1 is
redundant for the IDH2/3 interconversion of isocitrate to α-KG, its knockdown impacting
on insulin secretion suggests a non-redundant role in this pathway. Elevated expression
of IDH1, for example as a result of miR-23 diversion, could potentially provide the cell
with excess α-KG ultimately providing elevated NADH through TCA cycle processing.
However, the exclusive production of NADPH in the cytosol, through export of TCA
cycle intermediates via excess downstream sources, could possibly be playing a part in
signalling networks such as its interaction with dioxygenases and its role in oxidative
stress as IDH1 mutations have been associated with increased oxidative stress in gliomas
(Gilbert et al. 2014). Furthermore, IDH1 is one of the key genes found to be up-regulated
in breast cancer stem cells, which were identified to have lower levels of reactive oxygen
species (ROS) (Diehn et al. 2009). The overabundance of α-KG in addition to Fe2+ serve
Isocitrate α-ketoglutarate
NADP+ NADPH
IDH1
TCA
cycle Anaplerosis
316
as cofactors for the activity of histone demethylases such as Jumonji C ultimately
switching gene expression on (Horbinski 2013).
Prolyl-4-hydroxylase is an intracellular enzyme required for the synthesis of all types of
collagens and is stimulated by the levels of α-KG, potentially provided for by IDH1 in
miR-23 depleted CHO cells. The functional roles of prolyl-hydroxylases or 2-OG
oxygenases (2-OG is 2-oxoglutarate also known as α-ketoglutarate or α-KG) have a wide
range of functional roles including hypoxic response (oxygen sensing), nucleic acid repair
and modification, fatty acid metabolism and chromatin modification (Loenarz, Schofield
2008). Prolyl-hydroxylase catalyses the conversion of proline and α-KG into
hydroxyproline, succinate and CO2 in the presence of iron and dioxygen. Prolyl-4-
hydroxylase has been demonstrated to have a higher affinity of α-KG than other PDHs.
This high affinity would suggest that the cell would possess an inherent increased
sensitivity in responding to TCA cycle defects. It is possible that the α-KG being
produced exclusively in the cytosol by IDH1 is being sensed and processed by P4HA1
into its products hydoxyproline and succinate. Succinate could be shuttled into the
mitochondria where it serves as a substrate for succinate dehydrogenase (complex II)
resulting in the increased oxygen consumption observed at this complex. A study by Zhao
et al (Zhao et al. 2009) demonstrated that ectopic expression of an IDH1 mutant was
found to result in the inhibition of the activity of prolyl hydroxylase (PDH) which could
be restored by feeding cells with α-KG. Various pathways are under the influence of these
hydroxylases including collagen synthesis, histone modification and transcription. It is
estimated that there are over 60 α-KG-dependent dioxygenases (Rose et al. 2011)
suggesting that changes in IDH1 levels could potentially affect multiple pathways.
With the observation of glutaminase not being elevated in its translation due to miR-23
depletion, the increase in the glutamate concentration of 65% across all miR-23 depleted
clones could be as a result of the catapleurotic export of α-ketoglutarate due to elevated
IDH1 (increased synthesis of α-KG from isocitrate) and its conversion into glutamate
(Owen, Kalhan & Hanson 2002).
Malate dehydrogenase (MDH) catalyses the conversion of malate to oxaloacetate in a
NAD+-dependent manner generating the reducing equivalent NADH, contributing to the
electron transport chain. It is considered the second rate limiting step in the TCA cycle as
oxaloacetate is the terminal substrate required to produce citrate in the presence of acetyl-
CoA to start the cycle over again. MDH was found to be increased in its expression in
317
miR-23 depleted CHO-SEAP clones. This increase in MDH expression could be
providing surplus NADH reducing equivalent for electron donation at complex I thus
contributing to the elevated OXPHOS previously observed. As an aside, interference at
these rate limiting steps can be responsible for mitochondrial dysfunction due to
metabolite starvation such as the “Crabtree effect” whereby exclusive conversion of
glucose to lactate inhibits TCA cycle recycling by reducing Acetyl-CoA production from
pyruvate (Dell'Antone 2012) (Fig. 4.9).
318
Figure 4.9: The “Warburg” and “Crabtree” effect
Figure 4.9: The “Warburg effect” (Red Box) demonstrates the high consumption of glucose by
cancer/CHO cells with exclusive conversion to lactate. Although energetically inefficient,
intermediates provide key substrates for biomass production to actively proliferating cells. As
pyruvate is converted to lactate and not Acetyl-CoA, the TCA cycle is starved of essential starting
metabolites (Acetyl-CoA and oxaloacetate) causing TCA cycle weakening termed the “Crabtree
effects” (Green box). However, the overexpression of certain enzyme such as IDH1 and MDH, in
addition to the increased abundance of glutamate, can supplement and strengthen the TCA cycle by
providing key intermediates such as α-KG and encouraging forward reactions.
Overexpression of MDH II in CHO resulted in increases in intracellular ATP and NADH
with a corresponding 1.9-fold improvement in integral viable cell density (Chong et al.
2010). Growth improvement was not an observed benefit in our miR-23 depleted stable
Glucose Lactate
Warburg effect
Biomass production
Crabtree effect
Glutamate
319
SEAP cell line despite NADH concentration being correlated with cellular proliferation.
It would be interesting to assess miR-23 depleted clones for their metabolic profiles
including ATP, NADH, NADPH, as well as TCA intermediate metabolites such as
malate, α-KG, citrate and isocitrate.
4.3.5 Combating oxidative stress
Evidence suggests that protein folding and the generation of reactive oxygen species
(ROS) is a byproduct of protein oxidation in the ER. Additionally, one of the main cellular
organelles involved in the production of ROS is the mitochondria which, in a state of
chronic nutrient/energy overload, the flux of nutrients through the mitochondrial ETS can
result in enhanced ROS production and ultimately oxidative stress (Malhotra, Kaufman
2007). The observation of elevated transcription of the SEAP gene as well as increased
oxidative phosphorylation within miR-23 depleted clones, in addition to metabolic
enzymes of the TCA cycle fuelling OXPHOS, suggests that miR-23 sponge clones by
comparison to their control counterparts have a higher degree of oxidative stress.
Numerous proteins were found to be up-regulated that have been shown to respond to
oxidative stress including Thioredoxin reductase 1 (TXNRD1: 1.66-fold), Heme
oxygenase-1 (HMOX-1 or HO-1: 2.33-fold), glutathione-s-transferase Mu-1 and
glutathione-s-transferase Mu-6 (GSTM1/GSTM6 – 1.56/1.74-fold up-regulated
respectively).
Thioredoxin reductase-1 is a cytosolic antioxidant which is known to reduce various
cellular compounds including Thioredoxin using NADPH. Reduction of thioredoxin
(TXN) activates it allowing for the reduction of oxidised cysteins in cellular proteins and
scavenge peroxides by peroxidredoxins (PRDX), thus protecting the cell against oxidative
stress (Cadenas et al. 2010). Thioredoxin reductase does appear to be able to directly
remove ROS by reducing oxidants and hydrogen peroxide due to its highly reactive
selenocystein (Cenas et al. 2004). Activated thioredoxin has a redox-active cystein pair
through which it interacts with other proteins to regenerate proteins damaged by ROS for
example H2O2-inactivated GAPDH (Fernando et al. 1992). Thioredoxin is also a co-factor
for methionine sulfoxide reductase, whereby it can reduce methionine sulfoxide residues
in oxidised proteins caused by ROS (Watson et al. 2004). Thioredoxin reductase deficient
cells demonstrated a decreased basal mitochondrial oxygen consumption rates (Lopert,
320
Day & Patel 2012) suggesting its role in maintaining efficient mitochondrial function
under harsh oxidative stressful conditions.
Heme oxygenase 1 is a ubiquitous inducible cellular antioxidant stress protein that
catalyses the conversion of the pro-oxidant molecule heme into the products biliverdin,
iron and CO. HMOX1 is up-regulated in response to cellular stress such as hydrogen
peroxide. Mitochondrial localisation of HMOX1 has been demonstrated to confer
resistance to oxidative stress, correct mitochondrial dysfunction and inhibit apoptosis and
gastric injury (Bindu et al. 2011). Interestingly, Thioredoxin reductase has been
demonstrated to regulate the induction of heme-oxygenase-1 expression (Trigona et al.
2006) suggesting a synergistic link between both pathways and that the increased
expression of HMOX1 in miR-23 depleted CHO-SEAP cells could be a result of oxidative
stress and not as a function of miR-23 diversion.
The ubiquitin-proteasome pathway (UPP) is the primary cytosolic proteolytic machinery
for the selective degradation and removal of damaged proteins. Studies have
demonstrated that the 26S proteosome is highly susceptible to inactivation during
oxidative stress whereas the 20S proteasome is relatively stable under a variety of
oxidative stresses (Reinheckel et al. 1998). The 26S proteosome has been found to be
responsible for the unfolding and degradation of most normal ubiquitinylated proteins
(missense mutations) but shows poor selectively for oxidised proteins (Ding, Dimayuga
& Keller 2006). Unlike the 26S proteasome, the 20S exhibits a high degree of selectivity
for degrading oxidised or otherwise damaged proteins resulting from oxidative stress
(Pickering, Davies 2012). The 26S proteosome is composed of a 20S and two 19S
subunits. Cellular adaptation to oxidative stress has revealed that dissociation of the 26S
proteome liberates the oxidative-damaged specific 20S subunit as a means of combatting
potential cellular damage in a transcriptional-independent manner (Grune et al. 2011). It
is possible that the down-regulation of PSMD7 (26S proteosome non-ATPase regulatory
subunit 7-like) which is a component of the 19S subunit is reducing the association and
formation of the 26S proteosome thus providing increased availability of the 20S subunit
potentially providing assistance for mitochondrial/ER derived oxidative stress.
In addition to their role in cellular metabolism and other molecular pathways such as
biosynthesis, IDH1/2 have been observed to play a vital role in cell protection against a
variety of insults that produce oxidative stress (Chang, Xu & Shu 2011). Reduced
expression of both IDH1 and IDH2 was observed to result in higher levels of ROS and a
321
greater oxidative damage in response to an oxidative insult (Jo et al. 2001, Lee et al.
2002). Additionally, several reports have demonstrated that the overexpression of IDH1/2
can offer cellular protection against oxidative stress (Kim et al. 2007, Shin et al. 2004).
For examples, NADPH produced by IDH1/2 can be utilised by glutathione reductase to
convert the oxidised form of glutathione (GSSG) to the reduced form (GSH) which can
neutralise free radicals and ROS species (Lee et al. 2002).
Following from this, two glutathione-s-transferase of the Mu (µ) isoform (1 and 6) were
found to be up-regulated in miR-23 depleted CHO cells. Maintenance of the cellular
antioxidant glutathione (GSH) in different cellular compartments is under the critical
regulation of GSTs. GSTs function in cellular detoxification of endogenous toxic
metabolites, superoxide radicals and exogenous toxic chemicals (Raza 2011). Reports
have documented the presence of the mu class of GSTs in rat and mouse mitochondria in
addition to rat brain mitochondria (Addya et al. 1994, Raza et al. 2002, Bhagwat et al.
1998). The reduced expression of GSTM2 was observed to increase oxidative stress in
spontaneously hypertensive rats (Zhou et al. 2008). The observation of the elevated
abundance of these antioxidant proteins further suggests that cellular and molecular
pathways, mediating an enhanced SEAP productivity are inducing oxidative stress which
is being compensated for by numerous cellular responses. Glutathione (GSH) and
Thioredoxin are considered the most important out of the four antioxidant systems in the
cell. The observation of both of these systems potentially being overactive again suggest
that miR-23 depleted CHO cells are under oxidative stress which is being adequately dealt
with. Without measuring the internal levels of ROS, it is premature to assume that the
elevated expression of these oxidative stress responding proteins are activated as a result
of elevated mitochondrial function and SEAP production, in addition to their expression
(TXNDR1) being a result of miR-23 depletion or responsive to cellular cues.
4.4 Proteomic analysis of enriched mitochondria
Cell fractionation to isolate mitochondria and profile them directly not only serves to
identify low abundant proteins that would be otherwise over-shadowed by high abundant
proteins, but also serves to refine the localisation of previously identified proteins as
functioning specifically in the mitochondria (Drissi, Dubois & Boisvert 2013).
The previous link between miR-23 and cellular bioenergetics through glutaminolysis
(Gao et al. 2009) highlights its potential role in our enhanced mitochondrial function
322
through other pathways such as the various proteins mentioned previously. Another
mitochondrial-related gene found to be under the translational control of miR-23 is the
mitochondrial transcription factor A (TFAM) (Jiang et al. 2013b). This gene is involved
in the transcriptional initiation of the mitochondrial genome (Campbell, Kolesar &
Kaufman 2012) in addition to mtDNA packaging and copy number (Uchiumi, Kang
2012). Although several of these validated gene targets of miR-23 (GLS, TFAM and
XIAP) did not appear to be DE when label-free mass spectrometry was carried out, it is
possible that miR-23 can influence cellular metabolism through various molecular means.
However, given the previous association of high CHO cell specific productivity with
oxidative metabolism (Young 2013) and the ability of the endoplasmic reticulum to
increase energy provisions through calcium signalling (Kaufman, Malhotra 2014), it is
possible that this enhanced mitochondrial function is not causative but symptomatic of
the enhanced production phenotype of miR-23 sponge CHO-SEAP cells.
4.4.1 A deeper insight into mitochondrial function
Lactate dehydrogenase (LDH) directly catalyses the interconversion of pyruvate to lactate
with the concurrent interconversion of NAHD and NAD+. LDH-A was observed to be
down-regulated in miR-23 sponge clones by 1.3-fold. Although lactate is considered the
lesser of two evils when compared to ammonia (Kim do et al. 2013), the propensity of
CHO cells to act like some cancers with respect to the Warburg effect, excessive
accumulation of lactate is an issue. Inhibition of LDH-A has been demonstrated to
decrease glucose flux in the glycolytic pathway (Chen et al. 2001) and has further been
demonstrated in CHO, upon siRNA-mediated inhibition, to increase hTPO production by
2.2-fold without influencing growth or viability (Kim, Lee 2007). These authors
suggested that the viability benefit of reduced LDH-A activity may be more apparent
under fed-batch conditions whereby a high glucose feed would result in faster
accumulation of lactate. Although measurement of lactate levels would be required within
our miR-23 sponge clones, this could be a potential explanation for the extended viability
phase observed under fed-batch conditions that was not apparent when in batch culture.
It has been suggested that the increase in recombinant protein production may drive
pyruvate into the TCA cycle or divert glucose towards the pentose phosphate pathway
(Sengupta, Rose & Morgan 2011). With the observation of enhanced mitochondrial
323
function in miR-23 depleted clones, it is likely that oxidative metabolism is favoured. The
reduced expression of LDH-A could also make pyruvate transiently more abundant for
channelling into the TCA cycle (Fig. 4.9). With the observation that glycolytic
metabolism and oxidative metabolism being associated with growth and productivity,
respectively (Young 2013), the potential reduced glycolysis not impacting on miR-23
sponge clone growth is interesting. It has been reported that a 20% reduction in glucose
consumption in a CHO cell engineered with reduced LDH-A and enhanced Bcl2
expression, had no impact on growth (Jeon, Yu & Lee 2011). Another interesting report
demonstrated that the co-inhibition of LDH-A and pyruvate dehydrogenase kinase
(PDHK, the enzyme responsible for PDH inhibition) resulted in the 90% increase in IgG
specific productivity without any appreciable impact on CHO cell growth (Zhou et al.
2011a).
The enzyme pyruvate kinase (PKM) was found to be down-regulated in miR-23 sponge
clones by 1.9-fold. This enzyme is responsible for the generation of pyruvate from
phosoenolpyruvic acid and is a key regulator of the metabolic fate of the glycolytic
intermediates (Iqbal et al. 2014). Its expression has been demonstrated to promote the
Warburg effect and tumour growth (Sun et al. 2011). Down-regulation of this gene could
be mediating a reduced dependency on glycolytic metabolism in miR-23 depleted clones
starving the TCA cycle of its perquisite metabolite Acetyl-CoA. Acetyl-CoA catabolic
processes were also found to be enriched for in GO assessment. Oxidative metabolism is
possibly supplemented, as indicated by the enhanced mitochondrial function, through the
various isotypes of TCA cycle enzyme such as IDH1 and MDH1. This is irrespective of
the potential enhanced activity of various TCA cycle enzymes through calcium signalling
suggested by the increased expression of LETM1 and the energy requirements of the ER
(Fig. 4.1). However, without assessing the specific consumption rate of glucose, it would
be premature to conclude that glycolytic metabolism is being perturbed in our miR-23
sponge clones or if they are simply eliciting an enhanced mitochondrial function.
Both MDH isotypes were identified as being overexpressed in miR-23 depleted CHO
cells indicating that the isolation of mitochondria was successful in unveiling
mitochondrial specific proteins such as the mitochondrial specific MDH2. These
enzymes mediate the conversion of malate to oxaloacetate in addition to generating the
reducing equivalent NADH (Marín-García 2012)
. Overexpression of MDH2 in CHO cells has been demonstrated to increase intracellular
ATP and NADH levels by 1.9-fold contributing to an increase in IVCC (Chong et al.
324
2009b). Weakening of the TCA cycle has been observed due to the export of vital
intermediates such as α-ketoglutarate and malate out of the mitochondrial matrix
(Dell'Antone 2012). The constant supply of oxaloacetate via MDH1 and 2 may allow for
the efficient energy turn over as indicated when respiratory control ratios were calculated
for miR-23 sponge clones, as no bottleneck is provided in addition to the reduced LDH-
A activity potentially providing excess pyruvate. The restoration of oxaloacetate has been
observed to drive continued TCA cycle function (DeBerardinis et al. 2007). Another rate
limiting metabolite α-ketoglutarate could be in constant supply through the elevated
abundance of IDH1, identified in both whole cell and mitochondria enriched experiments.
This could potentially compensate for the observation of the reduced expression of the
mitochondrial-specific isotype, IDH3A, the predominant isotype involved in the TCA
cycle (Yang et al. 2012). It was also observed that other enzymes involved in the TCA
cycle were found to be down-regulated such as 2-oxoglutarate dehydrogenase (KGD1)
and aconitate hydrotase (ACO2). However, it is possible that the small changes in
abundance may not be impacting significantly on mitochondrial function as indicated in
intact and permeabilise assays.
Enzymes involved in combatting oxidative stress were found to be increased in miR-23
sponge clones including those previously identified in the whole lysate experiment such
as TXNDR1, HMOX1 and GSTM1/6. Mitochondrial Glutathione reductase (GSR) was
found to be upregulated 1.3-fold and serves as a key enzyme that converts oxidised
glutathione (GSSG) to its reduced form (GSH) which ultimately scavenges ROS (Gill et
al. 2013). Peroxiredoxin-2 (PRDX2) was also found to be increased in expression and has
been linked with protecting proteins against oxidative damage (Rhee et al. 2012). The
observation of the increase in expression of additional proteins involved in conferring
resistance to oxidative stress further suggest the increase in mitochondrial function
despite certain TCA cycle members being reduced.(Li et al. 2005)
Several proteins found to be associated with the electron transport system were found to
be DE in mitochondrial enriched miR-23 sponge clone extracts. Cytochrome c oxidase
(COX), also known as Complex IV, is the terminal enzyme complex of the mitochondrial
electron transport system that couples the transfer of electrons from reduced cytochrome
c to molecular oxygen and is composed of 13 subunits of both nuclear and mitochondrial
origin. The subunit COX5A, nuclear in origin, was found to be down-regulated in miR-
23 sponge clones by 1.8-fold. This complex has been associated with efficient assembly
325
of the COX supercomplex. The first step in COX assembly is the membrane integration
of COX1 followed by COX4-COX5A (Stiburek et al. 2005). Furthermore, direct depletion
of COX5A resulted in the decreased levels of assembly of the COX enzyme with a parallel
accumulation of COX sub-(Fornuskova et al. 2010)complexes and unassembled subunits
(Fornuskova et al. 2010). Another mitochondrial complex associated protein UQCRC1
(ubiquinol-cytochrome b-c1 reductase) was found to be down-regulated by 1.5-fold in
miR-23 sponge clones. This enzyme is responsible for transferring electrons form QH2 to
ferricytochrome c. Specific loss of UQCRC1 has been demonstrated to be induced upon
exposure to oxidative stress (Shibanuma et al. 2011). The reduction in abundance of both
these Complex III and IV related proteins could potentially explain the bottleneck in
electron transfer suggested earlier for the lack of synergy upon stimulation of Complex I
and II in permeabilised assays. Furthermore, stimulation of Complex IV by TMPD
demonstrated a reduction in oxygen consumption in miR-23 sponge clones potentially
due to the reduced expression of COX5A and associated assembly of COX. The continued
healthy phenotype of miR-23 deplete clones would suggest that the reduced expression
of these mitochondrial related proteins is not detrimental to the cell.
The protein NDUFA1 was once thought to be associated with complex I is now emerging
as having a role with complex IV (Carroll et al. 2006). NDUFA4 has been shown to be
associated with COX function but not in its assembly (Pitceathly et al. 2013) suggesting
that, despite its increased expression, will not enhance COX function. One observation
however indicates that enhanced mitochondrial function maybe due to the increased
abundance of metabolites providing excess quantities of the reducing equivalent NADH
and FADH2. This observation was the reduction in abundance of two components of
Complex II, Succinate dehydrogenase (SDH), A and B. SDH catalyses the oxidation of
succinate to fumarate in addition to the reduction of ubiquinone to ubiquinol (Kim et al.
2013). The dual role this enzyme plays links the oxidation of glucose in the TCA cycle
with the production of ATP. The reduced expression of two components does not
correlate the increased oxygen consumption previously observed at complex II during
permeabilised assays. However, this could be due to the mild reduction in abundance of
1.3 and 1.4 for SDHA and SDHB, respectively, potentially not abrogating enzymatic
function. Furthermore, the enhanced activity of other TCA cycle enzymes such as IDH
and MDH could provide excess metabolites mediating the observed enhanced activity.
326
The indication towards enhanced mitochondrial function with elevated SEAP
productivity as a result of stable miR-23 depletion begs the question: Is SEAP
productivity enhanced as a function of oxidative metabolism or is oxidative metabolism
elevated as a result of enhanced SEAP transcription and processing? Previous studies
have documented the ability of MYC to induce mitochondrial biogenesis, function and
enhanced oxygen consumption (Li et al. 2005). Metabolic reprogramming in activated B-
cells has also been recently reported with both an increase in glycolytic and oxidative
metabolism being observed (Le et al. 2012). The ability of MYC to regulate the
expression of several(O'Donnell et al. 2005b) miRNAs (O'Donnell et al. 2005b), miR-
23a/b included (Gao et al. 2009), including its role in metabolic reprogramming (Eilers,
Eisenman 2008) and the association of miR-23 with B-cell differentiation (Li et al.
2011a), further strengthens the potential of miR-23 playing a direct role in cellular
metabolism.
4.5 Potential bona-fide targets of miR-23
Identification of true miRNA targets is a cumbersome and trialling task, which without
appropriate validation, leaves the researcher with an extensive list of potential targets all
of which have question marks over them. Earlier, we demonstrated that by integrating
multiple “omics” datasets, low scoring in-silico predicted targets that would otherwise be
discounted can ultimately be discovered. However, when posed with a list of DE proteins
from a single dataset, prediction algorithms are still the number one source for narrowing
down the potential target list. Of the 41 protein targets found to be DE in the whole cell
lysates of stable miR-23 depleted clones, 13 were found to overlap with the list generated
from the miRWalk algorithm. Given that the miR-23 sponge clones divert miR-23 away
from its endogenous mRNA targets, mRNA transcripts under direct influence of miR-23
should demonstrate an increase in their level of translation. When this was considered, 8
protein targets (LETM1, CALM1, CaMKIIδ, FERMT2, IDH1, TXNRD1, PPP1CB and
P4HA1) of the 13 overlapping targets were found to be up-regulated. These eight targets
present themselves as priority targets of miR-23a/b. miRNAs impact on an endogenous
target is generally accepted to be modest, mediating an impact of ~ 2-fold
increase/decrease in the level of translation of the target protein (Meleady et al. 2012b).
All of these candidates demonstrate this modest change in translation ranging from 1.67-
2.73 with LETM1 exhibiting the highest level of translational de-repression. Of course
327
using the term de-repression is premature at this point as further validation would be
required as a means to assess the totality of miR-23s influence on these targets.
Detecting translational de-repression has been observed in the past to be easier than
translational inhibition despite both eliciting modest impacts on the translatome
(Martinez-Sanchez, Murphy 2013). From this perspective, it would be worth carrying out
label-free mass spectrometry on the parental CHO-SEAP cell line upon the transient
overexpression of miR-23 such as in the case of miR-7 overexpression with subsequent
proteomic analysis (Meleady et al. 2012a). Transient over-expression of miR-23a/b with
subsequent analysis of each of the above targets on both the mRNA and protein levels
would be the next step in elucidating their potential as true targets. Furthermore, siRNA
inhibition of each target in stable CHO clones engineered to deplete miR-23 would be a
means to determine which of the panel of targets is mediating the observed phenotype of
enhanced SEAP mRNA transcription contributing to elevated specific productivity and
those candidates whose expression is potentially symptomatic and not causative of the
induced phenotype. Lastly, cloning the 3’UTR of the highest priority targets downstream
of a reporter gene would be required to validate the potential suppressive role of miR-
23a/b.
LETM1 presents itself as the most attractive and high priority candidate in relation to
being a target of miR-23. As a means to account for variation between clones, all clones
were grouped together i.e. triplicate samples of miR-23 sponge clone 4 and 6 being
analysed as 6 miR-23 replicates. This would allow for the variation within clones
themselves to be accounted for with only protein candidates under the influence of miR-
23 demonstrating significant differences between the control groups. When replicates
from both miR-23 and miR-NC groups were randomised and compared, two proteins
were found to be DE that did not overlap with the final DE list. This was performed to
assess whether identifications were a result of random chance. LETM1 demonstrated no
DE between control groups while not exhibiting DE when both miR-23 sponge clones
were compared. To further refine the list and potentially eliminate clonal variation and
background noise, more miR-23 sponge clones and control clones could be selected and
assessed for differential protein expression.
328
4.6 Integrated “omics” profile of CHO cells with varying growth rates
The use of miRNAs in CHO cell engineering has flourished in the past few years and
have been secured in the toolbox of viable engineering strategies (reviewed recently in
(Jadhav et al. 2013)) and highlighted in the stable manipulation of miR-23 in CHO-SEAP
cells. Numerous DE profiles have been carried out to date in an attempt to identify
miRNAs involved in the regulation of bioprocess relevant phenotypes including
temperature shift (TS) (Gammell et al. 2007, Barron et al. 2011a), growth phases
(Hernandez Bort et al. 2012), apoptosis induction (Druz et al. 2011) and CHO cells grown
in serum-containing medium (Jadhav et al. 2013). These profiles have resulted in stable
CHO cell manipulation of candidates such as miR-7 (Sanchez et al. 2013b) and miR-466h
(Druz et al. 2013) enhancing bioprocess phenotypes.
With the advent of biphasic culture, accumulation of cell density is crucial before entering
the “production phase” whereby growth is compromised to the benefit of specific
productivity. We sought to potentially enhance growth early in culture using miRNAs
derived from this profile with an application in temperature shift culture, published in
BMC Genomics in 2012 (Clarke et al. 2012).
Additionally, by acquiring transcriptomic data, mRNA and protein, a systems level
perspective of the molecular networks surrounding CHO cell growth was revealed,
demonstrating gene ontology enrichment for cellular processes such as translation and
RNA splicing. Furthermore, the integration of all three datasets facilitated the
identification of potential mRNA targets under the direct regulatory control of the DE
miRNAs. This approach also highlighted targets that can be scored poorly by prediction
algorithms, Rab14-miR451 (Wang et al. 2011a), in addition to identifying “non-seed”
mRNA targets whose miRNA: mRNA interaction does not follow the binding criteria by
which most prediction algorithms are based upon (Pasquinelli 2012).
Given the large inherent genetic heterogeneity between CHO cell lines, demonstrated
when comparing the genomes of 6 CHO cell lines including CHO-K1, DG-44 and CHO-
S against the ancestral sequenced CHO-K1 genome (Lewis et al. 2013), it was essential
to minimise variation through the careful selection of CHO cells that would form the basis
of the DE profile. Chromosomal heterogeneity has also been observed even among cells
within a clonal population (Wurm, Hacker 2011), not to mention the genetic
329
heterogeneity associated with recombinant CHO cell line development (Matasci et al.
2011).
All things considered, the study sought to minimise variation through the careful selection
of mAb-secreting CHO cell lines that were derived from the same transfection pool,
exhibiting a range of growth rates spanning from 0.011-0.044 hr-1. To further eliminate
biological noise and expose variations related to the proliferation phenotype, sister clones
selected demonstrated a similar level of cell specific productivity of 24 pg/cell/day (±3),
an attribute that commonly varies with growth rate (Barron et al. 2011a)
4.6.1 Identification of miRNAs associated with cellular growth rate
Of the 51 DE miRNAs identified within this study, a number of them have previously
been associated with cell growth. Several miRNAs that were found to be upregulated
form part of a well-studied oncogenic cluster, miR-17~92 (e.g. miR-17, miR-18a, miR-
18b, miR-20a, miR-20b and miR-106a) (Olive, Li & He 2013). This cluster has been
shown to be directly activated by the oncogenic transcription factor Myc (Dews et al.
2006b) and E2F1 (O'Donnell et al. 2005a). Its repression has also been demonstrated to
be influenced by p53 (Yan et al. 2009b). Individual cluster components contribute to the
clusters pleiotropic function through the regulation of various phenotypes such as
proliferation (p21), apoptosis (Bim) and angiogenesis (Tsp1) (Olive, Jiang & He 2010a)
(Fig. 1.3 of the Introduction). This miRNA cluster has previously been identified in
CHO (Hernandez Bort et al. 2012). Functional redundancy has been suggested by Hackl
et al., by identifying the conservation of MREs within 26 validated targets of this cluster
through next-generation sequencing (such as E2F1, CCND1, BCL211 and PTEN) in CHO
(Hackl et al. 2011).
Among the 51 miRNAs DE within fast growing CHO cells, 15 make up the lesser
characterised miR-passenger (miR*) strands. Thought initially to be a by-product or
carrier (Guo, Lu 2010) of miRNA processing, it is now emerging that miRNA* strands
accumulate to substantial levels in-vivo, indicating a functional role (Chiang et al. 2010).
Bioinformatics’ analysis has demonstrated that a substantial fraction of miR* species are
conserved throughout vertebrate evolution with the greatest degree of conservation lying
within their seed regions and MREs (Yang et al. 2011). Growing evidence has supported
the functional role miR* strands play in genetic regulation such as target reporter gene
suppression as a function of endogenous miR: miR* reads (Yang et al. 2011), the
330
generation of mature miRNAs from both arms of a miRNA precursor (Okamura et al.
2008), the sporadic switch from the dominantly selected miRNA (Griffiths-Jones et al.
2011), the capacity of particular miRNA precursors to generate comparable levels of
mature miRNAs from both arms (miR-3p/5p) and ultimately their role in various disease
states such as arthritis (Niederer et al. 2012) and cancer (Chang et al. 2013).
13 of the 15 miR* strands DE in fast growing CHO cells were observed to be upregulated.
In some cases (miR-34a* and miR-23b*), this apparent increase in miR* strand
accumulation was not accompanied by an increase in transcription based-on the constant
levels of the mature guide miRNA strand. This could be as a result of a phenomenon
known as target-mediated miRNA protection (TMMP), a proposed mechanism by which
increased target availability can lead to the accumulation of miRNAs protected by
exoribonuclease turnover through target association (Chatterjee et al. 2011, Chatterjee,
Grosshans 2009).
miR-34a* has been observed to be down-regulated in Rheumatoid arthritis increasing
apoptosis resistance in synovial fibroblasts through the regulation of XIAP with its
overexpression inducing apoptosis (Niederer et al. 2012). This is in contrast to the
observation of its increased expression in fast growing CHO cells. However, the transient
overexpression of miR-34a* in CHO did correlate with this observation. Other miR*
strands over-represented are part of the well characterised miRNA clusters such as the
miR-17~92 and miR-23b~27b/miR-23a~27a clusters. There have been observations of
functional redundancy between partner strands through target sharing or pathway overlap.
One study demonstrated the induction of prostate tumour growth through co-ordinated
suppression of TIMP3 by both the guide miRNA, miR-17-5p, and the passenger miR-17-
3p (Yang et al. 2013). The miR: miR* relationship is further exemplified with the
observation that both miR-17-5p and miR-17-3p act synergistically to induce
hepatocellular carcinoma through the interference of independent signalling pathways via
knockdown of PTEN (-5p), GalNT7 and vimentin (-3p) (Shan et al. 2013). This would
suggest that miR* strands identified as DE could mimic their guide counterparts.
Several miRNAs found to be down-regulated as growth rate increased included miR-338-
3p, miR-204, miR-206 and miR-30e. One interesting observation was the down-
regulation of miR-30e in fast growing CHO cells and its up-regulation in CHO cells
subject to temperature shift growth arrest (Barron et al. 2011b). miR-451 has been shown
331
to inhibit growth (Li et al. 2013) and induce apoptosis (Bian et al. 2011) upon its
overexpression whereby it was observed to be down-regulated in fast growing CHO cells.
It is extensively reported within the literature that miRNAs are both species and cell
type/tissue specific. The nature of this variance is dependent on the availability of miRNA
targets and the sensitivity of the molecular pathways subjected to translational influence.
Several miRNAs play a recurring role within the literature, such as miR-34a and the
oncogenic cluster miR-17~92, earning themselves an almost p53-like status. From the list
of miRNAs found to be DE in our fast growing CHO cell phenotype, many correlated
well with reported function in the literature in addition to a subset of miRNAs previously
studied in CHO such as members of the miR-17~92 cluster (Jadhav et al. 2014) and miR-
30e (Barron et al. 2011a).
4.6.2 Transient screening of miRNA
Numerous miRNA profiling experiments have been carried out in CHO cells to include
subsequent screening strategies (Barron et al. 2011b, Jadhav et al. 2014, Strotbek et al.
2013, Jadhav et al. 2012). One such strategy began with an initial screen of 879 miRNA
mimics at a high concentration of 1 µM isolating candidates based on improved
volumetric productivity (Strotbek et al. 2013). Although this study identified that the
stable overexpression of miR-557/-1287 improved viable CHO cell density and specific
productivity, it highlights the potential loss of information because of the assessment of
a single phenotype (volumetric productivity) or the screening of only miRNA mimics.
Selection of the host cell line that will form the basis of a screening project will also
contribute to the outcome of the screen. This primarily is due to the variation in the
miRNA/mRNA expression across species, tissues and cell types. This is a critical factor
to consider when the cell system forming the basis of the DE profile is different to that of
the cell type exposed to transient evaluation. This potential variation was evident within
our primary screen with candidates such as miR-23b*, miR-34a*, miR-206 and miR-639
inducing a contrary phenotype.
The absence of phenotypic conservation within the CHO-SEAP screen could be due to
various reasons. One study demonstrated that predicted mRNA targets of highly
conserved tissue-specific miRNAs were expressed at low abundance in comparison to
tissues that did not express these tissue specific miRNAs. It also highlighted that highly
expressed genes were enriched that do not contain MREs within their 3’UTR for certain
332
tissue specific miRNAs (Sood et al. 2006). A study using a miRNA decoy library found
that over 60% of miRNAs had no discernible activity which indicated that the functional
“miRNome” may be quite limited and that only the most abundant miRNAs mediate
target suppression (Mullokandov et al. 2012). It has been identified that a “threshold”
relationship is apparent between miRNAs and their cognate mRNA targets (Pasquinelli
2012) and that a non-linear trend exists between a targets transcription and translation
(Mukherji et al. 2011). Published screens as mentioned previously have reported a small
numbers of impactful miRNAs which is in keeping with the results observed during
transient evaluation.
The fundamental point to take from these observations is that a miRNAs influence is
complex, to say the least. The observation that many miRNA candidates did not elicit the
anticipated phenotype as suggested from the profile does not rule out the potential for
these miRNA candidates to control CHO cell growth in a different CHO cell type. It is
possible that the molecular make-up of the CHO-K1-SEAP cell line used for transient
screening is different to that of the fast and slow CHO clones, thereby presenting a wealth
of differing molecular targets.
4.6.2.1 miR-23b*
Of the 15 miR passenger strands found to be DE in fast growing CHO cells, miR-23b*
exhibited the highest increased fold change of 121.Very little is known about the specific
role that miR-23b* plays. However, a recent report has identified it to be elevated in renal
cancers (Liu et al. 2010a) mediating its function through inhibition of the mitochondrial
tumour suppressor Proline Oxidase (POX) or alternatively Proline Dehydrogenase
(PRODH), correlating with its increased abundance in fast-growing CHO cells. Transient
overexpression, however, of miR-23b* demonstrated contrary results to both the
literature and the profile, with miR-23b* inhibition being demonstrated to elicit a similar
outcome (Liu et al. 2010a). A maximal inhibition of ~70% peak cell density was observed
upon miR-23b* overexpression with a mild induction of cell death. A more in-depth
analysis using the Guava Nexin assay was carried out to determine the degree of influence
this miRNA played in apoptosis. It was noted that there was an almost 2-fold increase of
cells entering early apoptosis (p ≤ 0.05) at both 48 and 96 h. The results are in contrast to
the reports above and suggests that the apparent increase in miR-23b* expression is a by-
333
product of the changing molecular environment and not a causative influence on the fast
growing phenotype.
One potential mechanism for the increased levels of this passenger strand in fast growing
CHO cells is the recently hypothesised target-mediated miRNA protection (TMMP).
Although the reciprocal relationship between miRNAs and their targets is not a new
concept (Pasquinelli 2012), TMMP is an experimentally validated mechanism by which
mRNAs, once thought to be passive targets of miRNAs, prevent the active turnover and
degradation of miRNA duplex components by cytoplasmic exoribonucleotlyic enzymes
such as XRN-2 in Caenorhabditis elegans (Chatterjee, Grosshans 2009) and XRN-1 in
humans (Bail et al. 2010) (Fig. 4.10). Target availability has been shown to be a
stabilizing factor for Ago-bound mature miRNAs (Chi et al. 2009). Chatterjee et al., has
demonstrated that target availability can also exhibit a protective effect on certain miR:
miR* strands (Chatterjee et al. 2011) (Fig 4.10 C). It potentially plays a role in RISC and
non-RISC bound miRNAs adding further complexity to the regulatory role that mRNAs
can elicit on endogenous miRNAs.
334
Figure 4.10: Enzyme-mediated microRNA turnover with miR* strand TMMP
Figure 4.10: A) Shows a pre-miRNA duplex with 3’ 2 nt overhangs being loaded into the Ago/RISC
protein complex with selection of the miRNA guide/passenger strand based on thermodynamic
stability of the 5’ end B) mRNA target binding by RISC directed by the miRNA guide strand with
exonuclease decay of the free passenger (miR*) strand by XRN-2 C) mRNA target availability
prevents decay of miR* strands through this TMMP process.
From the raw TLDA profiling data of the expression patterns of miR-23b and miR-23b*
(Appendix Fig. A10), it was observed that in slow growing CHO cells, a classical
expressional relationship between star and guide strand existed, one strand present at high
levels (miR-23b) and one strand at low levels (miR-23b*). However, in Fast-growing
cells, expression levels of both strands were comparable. It has been shown that some
miRNA precursors can be processed to produce significant amounts of mature miRNAs
from both arms (Okamura et al. 2008), in addition to arm-switching being tissue specific
(Griffiths-Jones et al. 2011). The increase in miR-23b* levels in fast growing CHO cells
versus slow with no associated change in the levels of mature miR-23b suggests that arm-
switching is not taking place in this instance. However, without assessing the
transcriptional levels of the precursor pre-miRNA, it cannot be concluded whether an
increase in transcription is compensating for the steady mature miR-23b levels or if arm-
Argonaute/RISC
Argonaute/RISC
A
XRN-2
miR*-passenger
miR-guide
5’-to-3’ exonuclease decay
B
C
miR*:mRNA TMMP
335
switching was taking place. A recent report has suggested that the abundance of target
transcript is primarily more selective at “protecting or mediating arm-switching” of miR*
strands over their more abundant guides (Kang et al. 2013b), perhaps due to the inherently
low miR* copy number. This dramatic impact on miR* strands was demonstrated by
Kang and colleagues whereby a 10-fold increase of miR-7b was observed compared to
~250-fold for miR-7b* in stable sponge models (Kang et al. 2013b). One major factor for
consideration from the perspective of “does the level of miR-23b* contribute to this fast
growing phenotype?” is whether the apparent elevated levels of miR-23b* are associated
with a functionally active RISC or are they merely a by-product of miRNA processing?
Stable depletion of miR-23b* in a fast growing clone from the profile was observed to
elicit a reduction in growth (Data not shown) suggesting that in this cell, miR-23b* may
contribute functionally.
From a productivity perspective, miR-23b* overexpression resulted in a maximum
reduction in SEAP production of 55%. This was attributed to the reduced cell density,
similar to transient miR-7 overexpression in the same CHO-K1-SEAP cell line (Barron
et al. 2011b). Although an increase in cell specific productivity is often observed upon
growth arrest, miR-23b*-mediated growth arrest did not generate such a phenotype. As
miR-23b* has been linked with the G2-phase of the cell cycle through POX interaction
(Liu et al. 2009), it is possible that no significant increase in specific productivity is a
result of the potential arrest in G2 whereas high specific productivity is often associated
with G1 phase arrest (Bi, Shuttleworth & Al-Rubeai 2004, Carvalhal, Marcelino &
Carrondo 2003).
Although miR-23b* overexpression did not present itself as an attractive option for CHO
cell engineering, a reversal of this observed phenotype would be desirable. Evidence
supporting reversing CHO cell phenotype was observed by Sanchez and colleagues
(Sanchez et al. 2013b) upon the stable depletion of miR-7 despite its overexpression
resulting in cell cycle arrest with reduced product yields (Sanchez et al. 2013a, Meleady
et al. 2012b, Barron et al. 2011c). Although little is known about the functionality of miR-
23b*, it is part of a well characterised miRNA cluster implicated in human disease, miR-
23a~27a~24-2 and its paralog miR-23b~27b~24-1 (Discussed briefly in the Introduction).
The pleiotropic functionality of this miRNA cluster is evident from the multiple processes
regulated by its members such as metabolism, oncogenesis, proliferation or
differentiation, with the additional observation from profiling studies in various disease
336
conditions that all three miRNA members are simultaneously deregulated (Chhabra,
Dubey & Saini 2010b). Individual members have been documented as potent modulators
of cellular processes such as miR-23b; found to regulate a plethora of phenotypes
including apoptosis, angiogenesis and invasion (Zhang et al. 2011). Interestingly, both
miR-23b and miR-23b* have been implicated in mitochondrial function, ultimately
contributing to metabolite introduction to the TCA cycle under the regulation of the
oncogenic c-MYC (Fig. 4.11).
In the case of miR-23a/b, c-MYC suppresses their expression thus derepressing inhibition
of glutaminase (GLS) and mediating the conversion of glutamine to glutamate (Dang
2010). In the case of miR-23b*, c-MYC enhances its expression levels thereby decreasing
POX expression (Liu et al. 2009). Genetic alternations linking the regulation of glucose
metabolism in cancers are evident by the aberrant expression of the oncogenic
transcription factor, c-myc, which directly regulates the expression of glucose metabolic
enzymes and additionally genes involved in mitochondrial biogenesis (Li et al. 2005,
Eilers, Eisenman 2008). Mitochondrial dysfunction has been observed in various
metabolic diseases including obesity, aging, diabetes and cancer, with miRNAs emerging
as key regulators of mitochondrial proteins encoded by nuclear genes (Tomasetti, Neuzil
& Dong 2013). Re-programming of mitochondrial metabolism to depend on glutamine
catabolism circumvents the requirement of glucose as a primary carbon source, starving
the TCA cycle of essential metabolites as observed via the “Warburg effect” (Dell'Antone
2012). Glutaminolysis, conversion of glutamine to glutamate, is mediated by the enzyme
glutaminase (GLS) of which miR-23a/b have been shown to target of whose transcription
is repressed by c-myc (Gao et al. 2009). The implication of miR-23a/b in cellular
bioenergetics encouraged the exploration of its stable inhibition in CHO cells.
337
Figure 4.11: Mitochondrial metabolic contribution of miR-23b and miR-23b*
Figure 4.11: The influence that miR-23a/b suppression plays in the release of Glutaminase under c-
MYC suppression thus alternatively feeding mitochondrial respiration through glutamate
metabolism independent of glucose. Additionally, as an alternative fuel source to glucose, miR-23b*
also contributes to mitochondrial function through proline metabolism. Figure was sourced and
modified from (Dang 2010) and (Liu et al. 2009).
4.6.2.2 miR-34a/34a*
miR-34a* was found to be elevated in fast growing CHO cells and displayed a similar but
less dramatic expression pattern with its partner strand miR-34a as miR-23b*
demonstrated with its counterpart. Given the extensive characterisation of miR-34a in the
literature, we thought it prudent to evaluate the phenotypic impact of both miRNAs on
the CHO cell despite only the star strand being DE. Transient transfection of either
miRNA resulted in a similar phenotype: i.e. detriment to cellular proliferation and
viability. This also supported a previous notion of miR:miR* components demonstrating
functional redundancy. miR-34a has been shown to be transcriptionally activated by the
“master tumour suppressor” p53 (Hermeking 2009), a transcription factor capable of
halting cell cycle progression or ultimately activating programmed cell death (PCD)
(Hermeking 2007). Ectopic expression of miR-34a has been observed to induce G1/S-
miR-23b*
338
phase arrest through direct inhibition of pro-cell cycle targets such as CCND1 and its
binding partner CDK6 (Sun et al. 2008b). Its counterpart on the other hand, miR-34a*, is
less well characterised and has been validated to target a member of the family of
apoptosis inhibitors, X-linked inhibitor of apoptosis (XIAP) (Niederer et al. 2012).
Reduced expression of XIAP has been correlated with a reduction in the level of
expression of CCND1 (Cao et al. 2013) inducing G1-phase arrest while another study
demonstrated that selective inhibition of XIAP arrested cells in S and G2/M-phase with an
associated deregulation of targets such as cyclin-B1 and CDK1/2 (Wang et al. 2013).
The small molecule inhibitor of XIAP, Embelin, induced the shift in expression of
apoptosis related proteins Bcl-1, Bax and Bcl-xL destabilising the mitochondrial
membrane potential allowing for the release of cytochrome c (Wang et al. 2013),
supporting the 30% reduction in cell viability reported for miR-34a* overexpression in
CHO in addition to a similar observation of its overexpression in synovial fibroblasts
(Niederer et al. 2012). Numerous anti-apoptotic genes have been validated within the
literature as direct targets of miR-34a including Bcl2 (Cole et al. 2008) and the inhibitor
of apoptosis (IAP) member, survivin (Shen et al. 2012) supporting its impact on CHO cell
viability, similar to miR-34a*. A recent study has identified XIAP to be a target of miR-
34a to add to the extensive list of validated targets (see table 4.2) (Truettner, Motti &
Dietrich 2013). On top of this, despite limited studies of miR-34a*, the realisation that
transient overexpression of both pre-miRs mediate similar phenotypes, it is possible that
both miRNAs share more common targets, beyond XIAP, associated with proliferation
and apoptosis than is currently known.
339
Table 4.2: Validated targets of miR-34a
Cellular
process
Gene Gene name Reference
Cell cycle
CDK4
CCNE2
CCND1
CDK6
Cyclin-dependent kinase 4
Cyclin E2
Cyclin D1
Cyclin-dependent kinase 6
(He et al. 2007a)
(He et al. 2007a)
(Sun et al. 2008b)
(Sun et al. 2008b)
Apoptosis/
p53 pathway
BCL2
SIRT1
YY1
BIRC5
B-cell leukemia/lymphoma 2
Sirtuin, silent information regulator 1
Ying yang 1 transcription factor (TF)
Survivin
(Cole et al. 2008)
(Yamakuchi, Ferlito &
Lowenstein 2008)
(Chen et al. 2011a)
(Shen et al. 2012)
Wnt
signalling/
metastasis
JAG1
WNT1
NOTCH1
LEF1
WNT3*
CTNNB1*
LRP6*
MTA2+
TPD52+
AXL+
Jagged 1
Wingless-related MMTV integration site member 1
Notch homolog 1, translocation-associated
Lymphoid enhancer binding factor 1
Wingless-type MMTV integration site member 3
Beta-catenin
Low density lipoprotein receptor-related protein 6
Metastasis associated 1 family member 2
Tumor protein D52
AXL receptor tyrosine kinase
(Hashimi et al. 2009)
(Hashimi et al. 2009)
(Du et al. 2012)
(Kim et al. 2011) *
(Kaller et al. 2011b) +
Cancer
MYCN
CD44
NANOG
SOX2
v-myc myelocytomatosis viral related oncogen N
Heparan sulphate proteoglycan
NANOG homeobox transcription factor
SRY (sex determining region Y)-box 2 TF
(Cole et al. 2008)
(Liu et al. 2011)
(Choi et al. 2011)
(Choi et al. 2011)
Mitotic
signalling
MET
MAP2K1
(Silber et
al. 2012)
RRAS
PDGFRA
Met proto-oncogen (hepatocyte growth factor receptor)
Mitogen-activated protein kinase 1 (MEK1)
Related RAS viral (r-ras) oncogen homolog
Platelet-derived growth factor receptor, alpha
(He et al. 2007c)
(Ichimura et al. 2010)
(Kaller et al. 2011b)
(Silber et al. 2012)
Oncogenic
Transcription
E2F3
MYB
MYC
E2F transcription factor 3
v-myb myeloblastosis viral oncogen homolog
v-myc myelocytomatosis viral oncogen homolog
(Welch, Chen &
Stallings 2007)
(Navarro et al. 2009)
(Christoffersen et al.
2010)
Metabolism
ACSL1
LDHA
IMPDH
Acyl-CoA synthetase long-chain family member
Lactate dehydrogenase A
IMP (inosine 5’-monophosphate) dehydrogenase
(Li et al. 2011b)
(Kaller et al. 2011b)
(Kim et al. 2012)
Post-
translational
modifications
FUT8
α-1,6-Fucosytransferase
(Bernardi et al. 2013a)
Table sourced and modified from (Bader 2012)
A maximum reduction of ~75-80% in SEAP activity was observed which was in parallel
with an almost 90% drop in maximal cell density upon overexpression of both miR-34a/-
34a*. As seen before, this decrease in yield is due to the limited number of actively
340
producing cells within the bioprocess. One interesting observation however was the
increase in normalised productivity for cultures treated with pre-miR-34a (section
3.1.2.3.1) compared to those treated with pre-miR-34a* (section 3.1.2.2.2). The
maximum increase of ~80% in normalised productivity in CHO-SEAP cells may be
perhaps due to the role that miR-34a plays in G1-phase arrest (Sun et al. 2008b) compared
to the association of miR-34a* with G2 (Wang et al. 2013).
4.6.2.3 miR-181d/200a
We pursued miR-200a to assess its possible functional conservation, low in relapsed
patients subsequent to prostatectomy (Barron et al. 2012), in CHO cells and observed a
reduction in cell density of 20-65% upon its overexpression (n = 7). This reduced cellular
proliferation coincided with improved normalised productivity, however, in some
instances volumetric SEAP yields were enhanced, suggesting that normalised
productivity was improved irrespective of growth. A study by Hennessy and colleagues
(Hennessy et al. 2010) found miR-200a to be down-regulated in glucose non-responsive
MIN6 cells. Transient inhibition of miR-200a decreased glucose-stimulated insulin
secretion (GSIS) which seems in keeping with the observation that its overexpression in
CHO cells potentially enhances SEAP production. The tumour suppressive nature of
miR-200a is exemplified in the study by Saydam and colleagues (Saydam et al. 2009)
whereby the expression of miR-200a was found to be ~25-fold down-regulated in
meningiomas with targets such as Cyclin-D1 and β-catenin (Su et al. 2012). Previous
characterisation of this miRNA to regulate CCND1 coincides with our observation of
reduced cell growth with enhanced normalised productivity (Kumar, Gammell & Clynes
2007). This would further suggest that miR-200a knockdown might improve CHO cell
growth thus allowing for the rapid accumulation of cellular biomass. Interestingly, miR-
200a was observed to be involved in a positive-feedback loop through the translational
inhibition of histone deacetylase 4 (HDAC4) in human hepatocellular carcinoma
implicating a role in transcription (Yuan et al. 2011). The specificity of HDACs is an area
of continuous research as to whether a set panel of genes or global gene expression is
influenced by their mechanism of action (Haberland, Montgomery & Olson 2009). The
potential of miR-200a to interfere with the translation of HDAC4 could potentially
contribute to the increased specific productivity independent of growth, if its panel of
targets excluded growth related genes for example.
341
The transient over-expression of miR-181d demonstrated a less potent and consistent
phenotype than miR-200a when screened in CHO-SEAP cells. Although selected for its
potential tumour suppressor influence, overexpression resulted in a varied suppressive
influence on CHO cell growth. One study reported that miR-181d, miR-34a and miR-
193b were down-regulated subsequent to MAPK activity (Ikeda et al. 2012). Although
miR-181d and the miR-181 family have been validated to target various cancer associated
genes such as k-ras, bcl2 (Wang et al. 2012), Mcl-1 (Ouyang et al. 2012) and Tcl1 in
chronic lymphocytic leukemia (Pekarsky et al. 2006), the expression profile of these
targets and fundamentally the retention of target conservation will dictate the potency of
this miRNA to elicit an effect in CHO. Of the two miRNAs screened, miR-181d and miR-
200a, it would be interesting to explore the potential synergy between the two however;
miR-200a would be a prime candidate for stable inhibition, in the future.
4.6.2.4 miR-532-5p
It has been suggested that lessons can be learned from nature such as mimicking the
differentiation of B-lymphocytes into plasma cells (Dinnis, James 2005) with a view to
potentially improving the secretory capacity of host rCHO cells. Numerous miRNAs have
been identified to be directly involved in glucose-stimulated insulin secretion (GSIS) such
as miR-375 (Poy et al. 2004) and miR-124a (Lovis, Gattesco & Regazzi 2008) in
pancreatic islet β-cells. miR-532-5p was found to be down-regulated in glucose non-
responsive MIN6 β-cells and was explored for its potential role in CHO cell productivity
(Hennessy et al. 2010).
Transient overexpression of miR-532-5p (Guide strand) had no impact on cellular
proliferation or viability but did impact negatively on specific productivity, ranging from
a ~20-30% reduction. Without assessing intracellular SEAP levels, it would be premature
to attribute this observed phenotype to a reduction in the secretory capacity of treated
cells or an impact on rSEAP transcription. In an attempt to reverse this phenotype, anti-
miRs designed against miR-532-5p were transiently transfected into CHO-K1-SEAP
cells with no impact on volumetric productivity. It was noted that the expression of miR-
532-5p in fast growing CHO cell clones was low (Appendix fig. A11) upon analysis of
the raw CT values from the original profiling study. It is possible that if miR-532-5p is
related to a fast growing phenotype then its low level of expression could be similar in
CHO-K1-SEAP. If that was the case, further inhibition may not yield a measurable
342
phenotype. However, upon exploring transient knockdown of this miRNA to mediate
enhanced SEAP secretion, there was a mild increase of around 20-30%, at its highest, in
cell numbers on day 2-4 of culture, suggesting that stable inhibition might be an
interesting route to pursue.
4.6.2.5 miR-639
miR-639 expression was increased in fast growing CHO cells, yet when over expressed
in our model CHO-K1-SEAP cell line, it caused a contrary but consistent reduction in
cellular growth with an associated increased normalised productivity. One study showed
miR-639 to be upregulated in the serum samples of patients with bladder cancer in an
attempt to evaluate the potential of miRNAs as candidates to assess prognosis and as
diagnostic biomarkers (Scheffer et al. 2012). This suggests it could act as an oncogene in
addition to a report of it targeting the tumour suppressor p21 (Wu et al. 2010). Despite
the suggestion from the literature, the transient overexpression of miR-639 induced a
reduction in CHO cell proliferation. It is possible that the increased expression of miR-
639 is a result of the phenotype and not a cause of it. The increase in normalised
productivity indicates that the observed reduced growth could be due to an arrest in G1
phase of the cell cycle potentially inducing an associated highly transcriptionally active
production phase (Kumar, Gammell & Clynes 2007). Despite its lack of characterisation,
miR-639 presents itself as an attractive candidate for stable depletion.
4.6.2.6 miR-206
Reduced expression of miR-206 in fast growing CHO cells agreed with numerous studies
reporting its decreased expression in gastric cancer samples and cell lines (Zhang et al.
2013), estrogen-dependent ERα-positive breast cancer (Kondo et al. 2008), melanoma
cell lines (Georgantas et al. 2014) and laryngeal cancer (Zhang et al. 2011). Several
characterised targets of this miRNA support its tumour suppressive capacity such as
CCND2 (Kondo et al. 2008), CCND1, CDK4 (Georgantas et al. 2014), VEGF (Zhang et
al. 2011) and Notch3 (Song, Zhang & Wang 2009). It was interesting to observe that
transient inhibition of miR-206 in CHO-K1-SEAP cells mediated a reduction in CHO cell
proliferation by 30% with an associated elevated normalised productivity of 15% on day
4 of culture.
343
The reason for its negative impact on CHO cell growth upon its transient inhibition
remains elusive. Reports of this miRNA in apoptosis suggests the potential for it to
prolong a viable production phase upon its inhibition (Wang et al. 2012). The observation
of enhanced normalised productivity suggests that cell cycle interference may be at the
G1 phase. It would be interesting to pursue its transient overexpression as a means of
assessing any conserved impact on CHO cell growth through targets such as CCND1/2
or CDK4.
4.6.3 Final miRNA screen comments
Several candidate miRNAs were highlighted from the transient screening process,
including miR-34a/34a*, miR-23b*, miR-200a, miR-206 and miR-639, some of which
were more potent than others and all of which demonstrated growth inhibition. The
observation that numerous miRNAs selected for screening either had no phenotypic
impact or the opposite of what was expected could be due to a point made earlier in that
over 60% of miRNAs expressed in any given cell type have no apparent functional
influence (Mullokandov et al. 2012). Additionally, the varied expression pattern of
miRNAs and mRNAs across various cell lines would suggest that miRNAs screened in
the CHO-K1-SEAP cell line could elicit a varied phenotype given the potential
differences in the expression profile of its mRNA targets. In essence, a miRNAs function
is dependent on the abundance and sensitivity of its gene targets.
The non-linear relationship between miRNA abundance and mRNA expression
(Mukherji et al. 2011) potentially contributes to a layer of complexity when assessing
miRNA function. When screening miRNAs in combination, a compromise has to be made
when selecting the concentration of each miRNA under investigation. Exogenously
introducing a high concentration, as in our transient screening process (50 nM), could
potentially saturate the miRNA biogenesis machinery when assessing multiple miRNAs
in combination leading to a potential transfection concentration of 150 nM and upward.
Aberrant expression of miRNA biogenesis components have been associated with
negative cellular phenotypes, for example Dicer knockdown inducing G1 arrest in MCF-
7 breast carcinoma cells (Bu et al. 2009) and the induction of mouse embryonic lethality
(Bernstein et al. 2003). We demonstrated that siRNA-mediated dicer knockdown was
detrimental to the CHO cell phenotype both inducing apoptosis and inhibiting cellular
proliferation, an observation also made by others in CHO (Hackl et al. 2014a). A CHO
344
specific observation was made that both Dicer expression and global miRNA abundance
were enhanced in CHO cells cultured in serum containing media (Hackl et al. 2011, Hackl
et al. 2014a). These observations not only exemplify the beneficial role that miRNAs play
in sustaining and maintaining cellular homeostasis but also that interference in global
miRNA processing can be detrimental to the cell. Compromising miRNA concentration
when screening multiples can potentially dilute miRNA activity to a point where a
phenotype is not perturbed which ties into this titrational effect that was previously
mentioned. Several miRNA candidates from the fast versus slow profile were assessed in
combination such as miR-338-3p, miR-455-3p/5p and miR-409-3p, members from the
panel screened that did not generate a significant/consistent phenotype. Additionally, in
combination with each other, a more consistent or potent phenotype was observed.
Not observing a phenotype upon transient manipulation of these miRNAs does not prove
that they do not play a role in the phenotype in which they were implicated. However, it
is possible that a large number of DE miRNAs in any profile are a result of the phenotype
and not the cause of it. The CHO-SEAP cell line is innately a fast growing cell line by
comparison to the fast growing clones used in the original profiling experiment. This
would suggest that a similar miRNA profile could be shared between both cell lines in
relation to growth. As a result, further manipulation of particular candidates may not elicit
a phenotype. Furthermore, the CHO-SEAP cell line could potentially demonstrate a
completely unique miRNA and mRNA profile thus not expressing targets of these
miRNAs. This suggests the benefit of profiling a panel of CHO cell lines from an
industrial point of view to determine the diversity of a particular targets across a range of
cell and product types. It would also suggest that profiling of candidates from a profile
should be carried out in the same cell line as used in the original profiling such as in the
case of miR-7 (Barron et al. 2011b) and miR-466 (Druz et al. 2011) in CHO.
By assessing and understanding the miRNA/mRNA expression patterns of CHO cell lines
two things can potentially be achieved. The first is the prediction of CHO cell
performance if miRNA/mRNA expression patterns are correlated with production
phenotypes such as growth or productivity. A study like this was carried out by Clarke
and colleagues which predicted CHO cell specific productivity to within 4.44 pg/cell/day
by using microarray technology generating a panel of 287 genes associated with
productivity (Clarke et al. 2011b). miRNA profiles across various CHO cell types, could
help to predict which CHO cells will respond to miRNA intervention.
345
Secondly, rudimentary inducible systems could be developed if miRNA expression
profiling data has been catalogued. For example, the cell cycle inhibitor p27 could be
engineered to include a miRNA sponge that contains MREs for a miRNA whose temporal
expression profile is confidently understood within a given CHO cell. If this miRNA has
a high expression early into culture then p27 will not be translated thus allowing for the
propagation of exponential growth (Fig. 1.7 of Introduction). If this miRNA is naturally
switched off or down-regulated in mid-late exponential growth then p27 translation
would be de-repressed inducing a growth arrest in G1 mediating an artificial, high
producing, temperature-shift like phenotype. This system could potentially provide
industry with an alternative to the use of exogenously inducible systems or the exploration
of temperature sensitive promoters. Profiling of tens or hundreds of CHO cell lines may
be costly but could present itself with numerous benefits such as those mentioned above.
The screening process also highlights the importance of selecting an appropriate cell line
for transient screening or suggests the necessity of carrying out screens in multiple cell
lines as a means of pin-pointing master miRNAs that possess conserved function across
various CHO cell types.
4.7 CHO cell engineering using miR-34
Our transient screening assay revealed potent suppression of CHO cell growth when miR-
34a was overexpressed. This observation prompted us to attempt a stable depletion
approach in anticipation of having the opposite outcome i.e. improved cell densities
and/or increased culture longevity.
4.7.1 miR-34’s depletion on various CHO cell types
The impact of a miRNA on a late stage culture phenotype such as apoptosis is difficult to
evaluate due to their short life-time. This can be caused by the stability of the exogenously
introduced miRNA in addition to cellular proliferation diluting cytoplasmic miRNAs
across daughter cells ultimately titrating their effect (Song et al. 2003, Bartlett, Davis
2006). To navigate around this, we cultured CHO-SEAP cells in nutrient-depleted media
and transiently inhibited miR-34a function resulting in a ~10% reduction in cell death.
This mild improvement in cell viability was similar to an observation by Druz and
colleagues upon transient inhibition of miR-466h in nutrient depleted cells (Druz et al.
2011). The inconsistent improvement in cell viability observed in the case of CHO-SEAP
346
cells could be attributed to the low endogenous levels of miR-34a in CHO cells reported
by next-generation sequencing (Hackl et al. 2011). Additionally, the successful profiling
of miR-466h by Druz and colleagues could be due to the same cell type being used in
both profiling and miRNA transient/stable inhibition (Druz et al. 2013). Transient
inhibition of miR-34a also demonstrated an inconsistent improvement of 30% in maximal
cell density. This increase in cell numbers could have contributed to the apparent increase
of 10% in culture viability.
The mild and inconsistent impact of miR-34a upon its inhibition may also be due to a
multitude of reasons: 1) the transient inhibition of miRNAs being less potent on
endogenous targets compared to their overexpression (Martinez-Sanchez, Murphy 2013),
2) the potential absence of molecular targets for miR-34a in combination with its low
abundance in CHO cells (Hackl et al. 2011) and 3) the dilution effect of miR-34a inhibitor
through multiple rounds of cellular proliferation. Stable inhibition of miR-34 was based
on previous observations in our group (Barron et al. 2011b), whereby stable depletion of
miR-7 enhanced cell density and boosted volumetric SEAP yield (Sanchez et al. 2013b)
despite increased apparent normalised productivity. Similar to miR-466h, miR-7 was
identified, characterised and stably engineered in the same cell line again highlighting the
selection of cell line having a contribution to outcome. Of course, from the perspective
of identifying a high priority miRNA candidate for industrial application, the effect
should be verified in several CHO cell lines.
When miR-34 depletion was assessed in a CHO-SEAP stable mixed population, there
was an increase in growth observed with an elevated viable cell density of 6-15% (Fig.
3.12 A and B). No benefit was observed in SEAP yield. However, when these cells were
cultured under fed-batch conditions, a considerable increase in SEAP yield was observed
(Fig. 3.13 C and D). This increase in productivity could have been attributed to the
reduced proliferation of miR-34 depleted cells when compared to controls, suggesting a
longer residence in the highly transcriptionally active G1-phase of the cell cycle (Kumar,
Gammell & Clynes 2007). When clones were assessed as a panel, volumetric SEAP
productivity was enhanced across all time points (Fig. 3.17). Cellular proliferation in 1
mL batch culture was reduced, possibly explaining the considerable increase in specific
productivity. Studies have demonstrated that CHO clones not only behave differently to
each other in respect to a certain phenotype but vary under differing bioprocess conditions
347
(Porter et al. 2010). Clones were selected based on their productivity and when cultured
in batch conditions continued to demonstrate an enhanced volumetric yield as observed
within the clonal screen. Interestingly, miR-34 clones exhibited a propensity to clump
when compared to controls and other miR-sponge clones generated (miR-23 and -23b*).
miR-34a has previously been associated with biological processes contributing to cell
adhesion (Jo et al. 2011), potentially explain this increased cellular interaction resulting
in clumping when clones were derived.
On the other hand, when miR-34 was stably depleted in IgG-secreting CHO-1.14 cells,
mixed populations grew better (Fig. 3.14 A). This agrees with numerous reports of loss
of miR-34a expression influencing cancer cell proliferation through mediating the
increased expression of numerous oncogenes such as CCND1, CCNE2 and CDK4/6 (He
et al. 2007b, Sun et al. 2008b). Loss of miR-34a function has been observed in a range of
cancers such as prostate (Fujita et al. 2008), breast (Vogt et al. 2011), pancreas (Chang et
al. 2007) and ovary (Corney et al. 2010) due to epigenetic mechanisms and genomic
instability (Calin et al. 2008, Vogt et al. 2011, Lodygin et al. 2008). The observation that
miR-34a overexpression reduced the mRNA levels of CCND1 in CHO suggested its
potential role in mediating a beneficial growth phenotype. However, the expression of
CCND1 in miR-34 depleted CHO clones varied considerably when compared to controls
(Fig. 3.11 C). It has been reported that mild changes of as little as 1.23-fold in the
expression levels of miR-34a targets have been sufficient to drive tumorigenesis (Kaller
et al. 2011a). It is possible that other targets both known and unknown are mediating this
enhanced proliferation such as CDK4/6. Additionally, siRNA knockdown of the miR-34
sponge did reduce growth suggesting that miR-34 depletion was eliciting an enhanced
growth phenotype.
Specific productivity was reduced across all time points in these mixed pool CHO-1.14
populations potentially as a result of enhanced proliferation. It has been suggested that
energy provisions can be redirected in fast growing cells from productivity to
proliferation and accumulation of biomass (Clarke et al. 2012). Furthermore, miR-34a
has been implicated in the arrest of cells in the G1-phase of the cell cycle upon its
expression (Navarro et al. 2009) ultimately prolonging their residence in this highly
transcriptionally active phase (Kumar, Gammell & Clynes 2007). Swift propagation
through the cell cycle could be contributing to the observed reduced specific productivity.
348
When miR-34 sponge clones were selected through FACS, the GFP expression pattern
across both CHO-SEAP and CHO-1.14 panels was suggestive that endogenous miR-34
was targeting the sponge as there was an average drop in GFP fluorescence across the
miR-34 sponge panels compared to controls, an observation made previously by Sanchez
and colleagues for miR-7 sponge clones (Sanchez et al. 2013b). Furthermore, it was
demonstrated that mature miR-34a levels when assessed in a set of miR-34a clones
exhibited a 5-fold reduction (Fig. 3.11 A). siRNA inhibition of the miR-34a sponge was
also shown to increase levels of endogenous miR-34a though not to resting levels when
compared to control sponges. This is possibly as a result of incomplete knockdown of the
sponge.
The enhanced growth phenotype observed in the case of CHO-1.14 mixed populations
was not retained when miR-34 clones were cultured in a 1 mL volume. However, once
introduced back into 5 mL volumes, miR-34 sponge clones showed an improved growth
phenotype over all control clones except one, which maintained high cell density and
viability. Volumetric IgG and specific productivity was observed to be reduced
considerably in the case of miR-34 sponge clones. This could be due to the more potent
increase in growth exerted by the miR-34 sponge thus contributing more significantly to
reduced specific productivity. When fed-batch was carried out, a similar trend was
observed in the case of miR-34 sponge clones, i.e. enhanced growth to the detriment of
productivity. However, individual miR-NC clones also performed well, showing high cell
densities and viability, miR-NC sponge clone 20, and high productivity, miR-NC sponge
2. This could be as a result of clonal selection, isolating individual clones that exhibit
desirable phenotypes potentially mediated by insertional mutagenesis resulting from
plasmid integrating or inherent heterogeneity.
From assessing miR-34a depletion across various culture formats within two cell lines, it
can be seen that it enhanced the specific production capacity of CHO-SEAP cells but
possibly caused these cells to clump mid-culture. This observation of miR-34 depletion
to improve productivity in selected clones suggests its potential use under an inducible
promoter, whereby a high specific production phase could be induced through depletion
of miR-34 mid-culture after biomass has accumulated. In the IgG-secreting CHO-1.14
cell line, miR-34 depletion potentially enhanced CHO cell growth at the expense of
productivity. However, the selection of a large clonal panel would be required before
concluding the miR-34 does confer a growth benefit to these cells. It is also possible that
349
in the case of the CHO-SEAP cell line, whose growth characteristics are far superior to
that of the CHO-1.14s, miR-34 depletion did not mediate an enhanced growth phenotype
as the potential molecular targets of miR-34 in this cell line would not mediate a further
improvement upon their deregulation. However, by investigating the potential role that
miR-34 could play in the bioprocess, the process of glyco-engineering would not have
been highlighted nor would the significant impact this engineering process would have
on industry be appreciated.
4.7.2 miRNA-based glycoform engineering of recombinant antibodies
It has been suggested that the maximum specific production capacity of CHO has been
achieved at 100 pg/cell/day (De Jesus, Wurm 2011), although not routinely accomplished.
An alternative means of indirectly increasing yield to meet the global demand for these
products could be offered through enhancing antibody effector function, ADCC, thus
reducing patient doses. The addition of a biantennary glycan (heptasaccharide) to the
conserved asparagine-297 considerably influences antibody effector functions (Jefferis
2009b). In fact, when the glycosylation is eliminated, binding of the target antigen has
been observed to decrease or not take place at all (Walsh, Jefferis 2006). The proportion
of non-fucosylated IgG-Fc produced in CHO is low with >90% of N-linked glycans
containing a core fucose group. It is possible that this high frequency of fucosylation
stems from their inability to add a bisecting N-acetylglucosamine as they do not express
despite containing a genomic homolog of the gene, β-1-4-N-
acetylglucosaminlytransferase III (GnTIII) (Xu et al. 2011). This was suggested by a
study whereby CHO cells engineered to express the GnTIII gene produced a non-
fucosylated Rituximab-like antibody product with enhanced ADCC due to prior addition
of a bisecting-GlcNAc (Ferrara et al. 2006).
shRNA technology designed against the FUT8 gene resulted in an 88% reduction in
fucosylation of an IGF-1R antibody with enhanced ADCC and little impact on
productivity in a CHO-DG44 clone (Beuger et al. 2009). Several studies have reported
improved ADCC function of CHO-produced antibodies for clinically relevant products
such as Herceptin and Rituxan exploring routes such as genetic knockout of FUT8
(Yamane-Ohnuki et al. 2004), GMD knockout (GDP-fucose 4, 6-deydratase (Kanda et al.
2007), siRNA-double knockdown of both FUT8/GMD (Imai-Nishiya et al. 2007) and
siRNA inhibition of FUT8 alone (Mori et al. 2004).
350
Having found that the stable depletion of miR-34 did not generally mediate a beneficial
phenotype, despite enhancing cell growth in the IgG-secreting CHO-1.14 cell line, we
explored its potential impact on the fucosylation of secreted antibodies following the
observation by Bernardi and colleagues that miR-34a could regulate the translation of
FUT8 (Bernardi et al. 2013a).
CHO cells produce predominantly fucosylated glycans (Kanda et al. 2006) with the main
glycoforms being GOF, G1F and G2F (Ucakturk 2012). It is possible that the increased
expression of FUT8, potentially as a result of miR-34 depletion, does not increase
antibody fucosylation. This is potentially due to the high abundance of fucosylated
antibodies (>90%) already produced by CHO. However, despite no change in
fucosylation being observed, there was a considerable reduction in peak intensity for the
miR-NC sponge clone when compared to miR-34 sponge clone possibly indicating a
reduced frequency in the post-translational addition of the N-Glycan at asparagine-297
(Fig. 3.27 in Results). Discrepancies have been observed in the glycosylation patterns of
clones derived from the same cell line indicating that clonal variation can contribute to
differences in post-translational modifications (Singh 2011). However, we were hopeful
that the depletion of miR-34 was enhancing the post-translational process occurring in
the CHO-1.14 cells. Little literature support could be found regarding the frequency at
which monoclonal antibodies are successfully glycosylated, however, we evaluated the
peak intensities of the parental CHO-1.14 cells and anticipated that a similarity would be
observed when compared to miR-NC sponge clones, i.e. reduced intensity by comparison
to miR-34 depleted clones. Both miR-34 sponge and parental CHO-1.14 demonstrated a
similar glyco-profile of comparable intensity (Fig. 3.29 in Results). Furthermore, siRNA
knockdown of the miR-34 sponge did not lower the peak intensity of the glyocoprofile
suggesting that the presence of the miR-34-specific sponge was not influencing the
frequency of N-glycan addition.
Interestingly, we observed no change in the glyocoprofile (Fig. 3.28 in Results) of the
miR-34 sponge clone 21 transfected with miR-34a mimic. This was carried out in an
effort to reduce FUT8 levels in the hope of reducing core fucosylation. In hindsight, this
was not the most ideal experimental setup as it is possible that the presence of a highly
expressed sponge specific for miR-34 could dilute miR-34a away from its endogenous
targets, including FUT8, in this instance (Mullokandov et al. 2012). When the miR-34
sponge clone 21 was transfected with miR-34a, there was a significant reduction in
351
maximum cell density observed (Data not shown). However, no decline in cellular
viability was observed, as had been previously induced in parental cell lines. This would
suggest that the presence of the miR-34 sponge competed for exogenously introduced
miR-34a preventing its suppression of genes involved in apoptosis resistance for example
Bcl2 (Yang et al. 2014) or XIAP (Truettner, Motti & Dietrich 2013). Although its potent
inhibition on cell growth was maintained, this could be due to the multitude of validated
targets of miR-34a involved in cellular proliferation such as CCND1, CCNE2, CDK4,
CDK6 and E2F1 thus hitting this phenotype from multiple angles. Transfection of miR-
34a into an IgG-secreting cell line not containing a sponge specific for miR-34a would
be a more desirable approach.
miR-34a function upon its overexpression in CHO appears to be conserved, as indicated
from the literature, in regards to reducing proliferation and inducing cell death (section
3.1.2.3.1). If its influence on FUT8 was also demonstrated to be conserved in CHO
resulting in reduced antibody fucosylation, as an engineering target it would still not
prove viable. As miR-34a overexpression would be required to reduce FUT8
translational, this could also result in it eliciting other phenotypes such as reduced cell
growth and cell death. As miRNAs are renowned for their ability to bind multiple mRNA
targets, it is possible that multiple miRNAs target the FUT8 gene. A method to identify
bound miRNAs to specific mRNAs has been reported and offers an attractive option to
achieve this goal (Hassan et al. 2013). By identifying novel miRNA targets of FUT8,
subsequent screening could identify those that mediate a potent inhibition of FUT8
translation without negatively impacting on other cellular processes such as growth or
apoptosis.
The benefit of exploring miRNA/siRNA inhibition of FUT8 in antibody producing CHO
cells would potentially navigate around the generation of multiple cell lines, i.e. the
primary or parental FUT8 knockout CHO cell lines and the subsequent introduction of
the specific antibody for large-scale production. A bi-cistronic expression vector has been
reported for the co-expression of the miR-183 family and a protein coding gene
(GFP/Atoh1) by exploiting the naturally occurring intronic pathway of miRNA
processing (Stoller, Chang & Fekete 2013). This technology would allow for the
simultaneous engineering of target therapeutic and miRNA. Although the generation of a
FUT8 genetic knockout cell line, as demonstrated by complete de-fucosylation of a
Rituxan antibody with enhanced ADCC (Yamane-Ohnuki et al. 2004), would be an
352
attractive option for industrial implementation, the availability of a co-expression system
could accommodate engineering of the multiple cell types suited for particular
bioprocesses and certain “hard-to-process” recombinant proteins. Glycoform engineering
offers industry a new option to produce highly active, less immunogenic and controlled
(Mode-of-action executed to suit treatment) antibodies that can indirectly address the
strain of global demand. As mentioned previously, current blockbuster antibodies such as
Herceptin and Rituxan will more than likely not be manufactured in these FUT8
engineered cell lines due to the costs involved in characterising their new structure in
clinical testing. The technology will, however, provide a viable platform for the
production and large-scale manufacture of the 350 antibodies currently in development
(Reichert 2013). Additionally, as patents expire, new parental cell lines will have to be
generated allowing for the introduction and implementation of this highly beneficial
engineering strategy. In addition, increasing the effector function of current antibodies
could also mediate the development of “Bio-betters”.
4.8 CHO cell engineering using miR-23b*
Previous characterisation of this miRNA in CHO-K1-SEAP cells demonstrated it to be
detrimental to the CHO cell phenotype upon its overexpression, resulting in reduced
maximum cell densities and the induction of apoptosis. This observation was in contrast
to that indicated from the profile. As previously suggested, the phenomenon of target-
mediated miRNA protection (TMMP) (Chatterjee et al. 2011) is potentially causing the
apparent increase in mature miR-23b* although the nature of it exerting a phenotypic
influence remains in question.
When miR-23b* levels were measured in miR-23b* sponge clones, there was no apparent
change when compared to non-specific controls (Appendix Fig. A13). This was in
contrast to the reduced levels of endogenous miR-34a and miR-23b in their respective
sponge clones as well as reports of miR-7 being up in its sponge cell line (Sanchez et al.
2013b). Despite an observed phenotype upon miR-23b* depletion and the demonstration
of the sponges specificity for miR-23b* via GFP inhibition, it was interesting to observe
no influence on endogenous miR-23b* levels. It has been demonstrated in HeLa cells that
miRNAs are expressed in a 13-fold excess relative to that of Ago1-4 indicating that there
are more freely available miRNAs than RISC-bound. Additionally, there is a 7-fold
excess of free miRNAs associated with mRNA targets compared to Ago-bound miRNAs
353
demonstrating that free miRNAs can associate with mRNAs and do so in excess to RISC-
bound miRNAs (Janas et al. 2012). This would also suggest that free miRNAs can
compete for and interfere with the regulatory capacity of their RISC-bound counterparts
also suggesting that TMMP can influence any miRNA within the transcriptome. On the
other hand, if there are more RISC-bound miRNAs than free, the interaction with mRNAs
could accelerate the turnover of these RISC-bound miRNAs upon dissociation from the
mRNA target. The fate of the miRNA-RISC complex is not something reported on in the
literature so the reduction in miR-23b/-34a levels through this increased instability upon
mRNA dissociation is purely speculative. By sequentially assessing numerous miR-
sponge constructs specific to miR-34a/23/23b* with various mismatches within the
MREs, it could be possible to evaluate if finite changes in complementarity will dictate
the observation of TMMP or reduced mature miR levels. The above questions the
influence of TMMP in both influencing apparent miRNA abundance and additionally in
the regulation of a miRNAs phenotypic influence.
miR-23b* depletion in CHO-SEAP mixed pools enhanced growth in both batch and fed-
batch conditions at the cost of specific productivity, resulting ultimately in reduced
volumetric SEAP output. This phenotype was conserved when assessed in the IgG-
secreting CHO-1.14 cell line with the addition of improved cell viability. Only one
validated target of miR-23b* has been reported, Proline Oxidase (POX) or Proline
Dehydrogenase (PRODH) (Liu et al. 2010a). This is a mitochondrial inner-membrane
tumour suppressor gene involved in proline catabolism. The absence or reduction of
PRODH has been observed in a variety of human tumours (Liu et al. 2010a, Liu et al.
2009) with its induction being observed to be regulated by p53 (Polyak et al. 1997)
resulting in cell cycle arrest in G2-M (Donald et al. 2001) and the initiation of apoptosis
through its ability to generate ROS (Liu et al. 2006). In miR-23b* depleted clones, the
enhanced growth was potentially associated with an increase in the target PRODH,
contradicting the reports mentioned. It is possible that the enhanced growth phenotype is
being mediated by a variety of potential miR-23b* targets that have not yet been identified
due to the little attention previously paid to miR* strands. Enhanced growth was not
observed when clones were assessed in a 1 mL volume, an observation made in the case
of miR-23 depleted CHO-SEAP clones also. As specific productivity and cell density
appeared to be equivalent on day 2 of culture, it is possible that the reduced cell numbers
observed on day 4 contributed to the increased specific productivity. However, in the case
of the IgG-secreting CHO-1.14 cells (mixed pool, clonal panel and clones), cell density
354
was enhanced at the cost of productivity. The phenotype mediated by miR-23b* was
observed to be retained in the CHO-1.14 cell line.
Upon selection of miR-23b* depleted CHO-SEAP clones, a range of growth phenotypes
were observed with clone 18 demonstrating the highest cell density and culture longevity.
miR-23b* sponge clones demonstrated a higher productivity then previously documented
for mixed pools but still lower when compared to controls. This potentially could have
resulted from the selection of clones based on productivity values thus introducing a bias.
Clone 18 produced a comparative amount of SEAP on D6 of batch culture despite its
reduced specific productivity. This may have been as a consequence of extended viability
towards the latter part of the culture compared to controls. This comparative productivity
was retained under fed-batch conditions with viability being maintained above 80% for a
further 72 h. Temperature shift was explored with the aim of improving specific
productivity and further delaying the induction of apoptosis which ultimately resulted in
improved SEAP yield of 50%.
One observation was that, upon subsequent experimental repeat of miR-23b* sponge
clone 18 in CHO-SEAP cells in fed-batch conditions including a temperature shift, a
reduction in GFP expression was observed which coincided with a decrease in viability
at an earlier time point in culture (Data not shown). This potentially implicates the
presence of the sponge in mediating the observed phenotype. Given the variation
observed within clones themselves over several generations (Pilbrough, Munro & Gray
2009), this reduction in GFP expression could also be attributed to an increase in
endogenous miR-23b* expression. Of the three miR-23b* sponge CHO-SEAP clones
selected from the panel screen, clone 18 demonstrated the lowest level of GFP expression
with clone 8 being the highest and clone 3 falling in between. Only clone 18 demonstrated
enhanced viability in 5 mL scale up suggesting that the presence of the sponge is not
influencing extended viability and may be a result of mutagenesis in addition to the
influence of the culture format (Porter et al. 2010). This type of forward genetic
characterisation has been explored to identify genes involved in cancer onset by using a
transposon, Sleeping Beauty (Dupuy, Jenkins & Copeland 2006) supporting the
possibility of beneficial CHO phenotypes induced by plasmid integration. No viability
benefit was observed across the panel of miR-23b* sponge clones indicating that this is a
phenotype not effected by miR-23b* depletion. Selection of a larger number of clones
355
carried over into the various culture formats would have given a better indication as to
whether this prolonged culture longevity observed for clone 18 was more than a by-
product of the cloning process. Regardless, the phenotype observed in miR-23b* sponge
clone 18 would be desirable under perfusion culture conditions. This bioprocess
configuration has been shown to compensate for low specific activity, as in the case of
miR-23b* depletion, due to the maintenance of high viable cell densities (Clincke et al.
2013) and demonstrates the utility of process selection for exploiting innate CHO cell
characteristics.
Although it would be pre-mature to conclude that miR-23b* depletion is mediating the
observed phenotype, it is evident that by selecting and adapting particular bioprocess
regimes to suit a clonal attributes, a desirable outcome can be achieved.
4.9 miRNAs place in the bioprocess
Although the specific production capacity of CHO cells has been proposed to have
reached its natural limit at 100 pg/cell/day (De Jesus, Wurm 2011), acquisition of this is
not routinely achieved and further drives the requirement for alternative engineering
approaches. One such alternative is boosting CHO cell densities and/or prolonging culture
viability without compromising specific productivity. It is for this reason that miRNA
profiling of CHO cells with varying growth rates but similar specific productivities was
explored (Clarke et al. 2012) identifying miRNAs associated with growth such as the
miR-17~92 cluster (Olive, Li & He 2013). This miRNA has already demonstrated
promising results in the manipulation of the CHO cell phenotype including the
overexpression of miR-17 enhancing growth performance, increasing specific
productivity 2-fold and boosting yield 3-fold of an EPO-Fc protein (Jadhav et al. 2014).
More recently, miR-17, -19b and -92a overexpression was observed to enhance antibody
titres by 130-140% in clonal cells without interfering with product quality (Loh et al.
2014). Other reports have demonstrated beneficial outcomes as a function of miRNA
intervention including miR-7 depletion (Sanchez et al. 2013b), miR-466h inhibition
(Druz et al. 2013) and miR-557/-1287 overexpression (Strotbek et al. 2013) all resulting
in enhanced product titres in SEAP and IgG producing cell lines. Characterisation of the
benefits of miRNAs over a range of recombinant products further enhances their utility
as engineering targets. Furthermore, several of the above reports demonstrate the
356
improvement in cell growth without compromising specific productivity, a common
trade-off often reported (Lai, Yang & Ng 2013a).
To add to this list, miR-23 depletion in CHO-SEAP cells was observed to enhance
specific productivity ~3-fold without impacting cell growth. However, pursuit of
inhibition of the entire miR-23~27~24 cluster could offer a more adventitious phenotype,
given its association with B-cell differentiation, and preliminary data in mixed pools
demonstrating both enhanced growth and specific SEAP productivity.
In the 7 years since the first miRNA differential expression profile of CHO cells subjected
to temperature shift (Gammell et al. 2007), numerous reviews have been published that
have documented the progress in the field and further speculating on miRNAs future in
the bioprocess (Wu 2009b, Jadhav et al. 2013, Muller, Katinger & Grillari 2008, Barron
et al. 2011d). Significant efforts have been invested to the understanding and control of
CHO cells and the molecular tools available for their engineering including the release of
the CHO genome (Xu et al. 2011), the utility of endogenous CHO miRNA clusters
(Klanert et al. 2013) and the large-scale transfection of CHO cultures with miRNAs
(Fischer et al. 2013) (Fig. 1.4 of the introduction).
Furthermore, the ongoing profiling strategies (Barron et al. 2011b, Hernandez Bort et al.
2012, Druz et al. 2011) in CHO have been suggested to benefit the bioprocess in the form
of inducible systems dependent on the innate expression profile of miRNAs over the
culture process (Fig. 1.7).
The possibility for miRNAs potential in bioprocess engineering does not end here. An
exciting area of research has been in the manipulation and control of particular antibody
glycoforms produced by CHO cells with the goal of enhancing antibody effector
functions such as ADCC (Jefferis 2009b). We explored the potential of miR-34a to retain
its function of suppressing the translation of the FUT8 gene (Bernardi et al. 2013b) as a
means of reducing the level of fucosylated IgG antibody. Although we were unsuccessful
in demonstrating miR-34a to impact on antibody fucosylation, this particular area of
research will have vast implications for the industry. By enhancing antibody effector
functions, therapeutic dosages can be reduced which could ultimately aid in achieving
global demands. This indirect route of meeting demand, or bringing the demand down,
will reduce costs involved in bioprocess design, allow for the derivation away from the
construction of large-scale pharmaceutical production plants and ultimately reduce
consumer costs.
357
With patents of major blockbuster therapeutics coming to an end, the drive for the
development of “bio-similars” and “bio-betters” is in full swing with over 350 antibody
products alone currently in pre-clinical development (Reichert 2013), a new era of cell
line development will be opened allowing for the re-optimisation of the parental
production cell lines for industrial scale manufacture. The various success stories reported
for miRNAs, miR-23 included, effectively enhancing CHO cell performance across a
range of recombinant products, implicates these small molecules as additional tools to
compliment other approaches.
358
Conclusions
1 Integrated microRNA profile of CHO cells with varying degrees of growth rates
and subsequent phenotypic characterisation through transient screening
51 miRNAs were found to be differentially expressed (DE) demonstrating a
positive or negative correlation with increasing CHO cell growth rate such as the
oncogenic miR-17~92 cluster.
Transient screening revealed miR-34a/34a* and miR-23b* to negatively impact
cell growth and viability: miR-206, miR-639 and miR-200a to negatively impact
growth with potential improved productivity in the case of miR-200a: and miR-
532-5p and miR-15a~16-1 to positively influence CHO cell growth.
Target-mediated miRNA protection (TMMP) could influence and reshape the
apparent miRNA transcriptome ultimately contributing to a symptomatic change
in mature miRNA levels, as indicated in the case of miR-23b*.
2 miR-34a as a potential candidate for CHO cell engineering
miR-34a overexpression was demonstrated to arrest cellular proliferation and to
induce cell death with concomitant improved specific SEAP productivity.
miR-34a was observed to enhance CHO cell specific productivity when stably
depleted in CHO-K1-SEAP cells but appeared to cause clones to clump mid-late
exponential phase of culture with clones demonstrating reduced culture viability.
In IgG-secreting CHO-1.14 clones, miR-34 depletion was observed to enhance
CHO cell growth at the expense of specific productivity potentially through the
regulation of targets such as CCND1.
Stable depletion of miR-34 did not appear to further increase antibody
fucosylation potentially attributed to the inherently high (>90%) level of
fucosylated IgG secreted by CHO.
Overexpression of miR-34a did not reduce the level of antibody fucosylation
potentially resulting from the presence of the miR-34 sponge titrating miR-34a
function away from FUT8.
3 miR-23b* as a potential candidate for CHO cell engineering
The over-expression of miR-23b* inhibited cell growth and induced cell death in
CHO-SEAP cells.
Stable depletion of miR-23b* appeared to enhance CHO cell growth at the
expense of productivity in both CHO-SEAP and CHO-1.14 cells.
In a stable CHO-SEAP clone, miR-23b* sponge clone 18, was observed to
produce ~50% more SEAP in a fed-batch temperature shift, despite its reduced
359
specific productivity compared to controls. This enhanced phenotype was
attributed to a prolonged production phase due to its enhanced culture longevity.
4 miR-23 depletion as an attractive tool for industrial application
miR-23 depletion using a miRNA sponge enhanced CHO-SEAP cell productivity
by an average of ~3-fold without affecting cell growth.
Productivity was demonstrated to be enhanced at the transcriptional level with no
influence on post-translational/secretory bottlenecks and mRNA stability.
Stable depletion of miR-23 in IgG-secreting CHO-1.14 cells enhanced cell growth
and appeared to prolong viability at the expense of productivity.
The variation of miR-23’s impact on two separate CHO cell lines, both derived
from the same parent, suggests that this variation is product-specific.
Mitochondrial function appeared to be enhanced in miR-23 depleted CHO-SEAP
clones.
Proteomic profiling of whole cell lysates identified a high priority target LETM1
(associated with mitochondrial function) to be elevated in miR-23 sponge clones
by 2.7-fold.
Elevated SEAP transcription is enhanced in miR-23 sponge clones possibly due
to the increased expression of CaMKIIδ and 14-3-3η which have been
demonstrated to interact with HDAC4/5 sequestering them in the cytoplasm
resulting in elevated transcriptional activation through reduced deacetylation in a
co-factor dependent manner.
Whole cell proteomic analysis further identified 8 potential targets of miR-23
which were found to overlap with predicted targets using the miRWalk algorithm
including LETM1, CALD1, CaMKIIδ, FERMT2, IDH1, TXNDR1, PPP1CB and
P4HA1.
Several proteins found to be enriched in both whole cell lysate and mitochondrial
extracts have been previously associated with cellular bioenergetics and
mitochondrial metabolism including MDH, SDH, GAA and IDH1 in addition to
proteins involved in oxidative stress such as TXNRD1, GSTM1/6, IDH1, PSMD7
and HMOX1 further indicating that CHO cell mitochondrial function is
potentially elevated as indicated through miR-23 depletion.
360
Future work
1 Identify additional miRNA targets for stable CHO cell engineering
It would be interesting to screen the full panel of DE miRNAs from the profile in
both a slow growing clone and the model CHO-SEAP cell line.
Explore the potential of TMMP playing a role in the apparent increase of miR-
23b* levels in fast growing CHO clones and if miR-23b* is itself influencing this
fast growing phenotype.
2 Validate the selectivity for enhanced SEAP transcription in CHO-SEAP cells upon
stable depletion of miR-23
siRNA-mediated inhibition of a selection of potential targets of miR-23 derived
from label-free MS such as LETM1, CaMKIIδ and IDH1 including other
interesting targets such as 14-3-3-η would assess whether these targets are playing
a role in the enhanced SEAP productivity phenotype.
Immunoprecipitation of candidates such as 14-3-3-η and LETM1 followed by
label-free mass spectrometry would give an indication of the cohort of proteins
associated with these upregulated targets such as the hypothesised family of
HDACs.
Over-expression of miR-23a/b in CHO-SEAP cells and western blot analysis of
the 8 high priority predicted targets (LETM1, CADL1, CaMKIIδ, FERMT2,
IDH1, TXNDR1, PPP1CB and P4HA1) would further prioritize their candidacy
as targets of miR-23 with the expectation of reduced translation. Following this,
3’UTR cloning of miR-23 responsive targets placed downstream of a GFP
reporter would confirm them as bona fide targets upon miR-23 overexpression.
Assess transcript levels of IgG (heavy/light chain) in miR-23-depletd CHO-1.14
clones to determine enhanced transcription with a potential down-stream
processing bottleneck.
3 Investigate the impact miR-23 depletion is exerting on mitochondrial function
Select a larger number of miR-23 and miR-NC sponge clones and assess
mitochondrial function using the OROBOROS under both intact and
permeabilised conditions. Furthermore, it would be interesting to assess variation
within clonal groups for differences in mitochondrial function.
siRNA-mediated sponge knockdown of both miR-23 and miR-NC sponges within
selected clones and subsequent mitochondrial function characterisation using the
OROBOROS.
siRNA-mediated knockdown of the SEAP transcript with subsequent
OROBOROS assessment could potentially determine if protein processing is
driving energy demand.
361
Carry out the quantification of the various metabolites such as isocitrate, α-
ketoglutarate, malate, oxaloacetate, lactate and ammonia in miR-23 sponge CHO-
SEAP clones versus controls. Additionally, determining the availability of NADH
and FADH2 could indicate the direction of metabolic reactions. Ultimately,
evaluating the levels of ATP will provide insight as to whether the enhanced
mitochondrial function is contributing to an increased energy provision.
4 Investigation of the role of miR-23~27~24 depletion in a panel of CHO cells
producing a range of recombinant products
With the previous observation that depletion of the entire miR-23~27~24 cluster
enhanced both cell growth and productivity without compromising specific SEAP
productivity, stable depletion and generation of CHO-SEAP (various recombinant
products) clones exhibiting stable cluster diversion could further enhance
volumetric titres due to increased cell densities.
5 Further explore the manipulation of antibody glycoforms using miRNAs
Assess the capability of miR-34a to interfere with the fucosylation of antibodies
produced in rCHO cells. Transiently transfect antibody-secreting cells with miR-
34a followed by western blot analysis of the α-1,6-fucosyltransferase (FUT8) to
determine effective targeting of FUT8 itself. Following this, glyco-profiling will
be explored to confirm reduced core fucosylation.
As miR-34a overexpression does not present itself as a viable option for CHO cell
engineering given its widely conserved phenotypic influence of reduced cell
growth and induction of apoptosis, additional miRNAs that target FUT8 will be
explored. The use of biotin-labelled anti-sense DNA oligonucleotides will be used
to pull down FUT8 with any associated miRNAs followed by miRNA target
identification using miRNA arrays.
From a practical application perspective, identified miRNAs will be validated as
bona fide targets of FUT8 in addition to interfering with core fucosylation.
Furthermore, overexpression of these miRNAs will be screened for any potential
adverse effects such as a negative impact on cell growth, viability and/or
productivity.
Finally, ADCC studies with the antibody products of these stable engineered
miRNA cells will be essential to confirm the enhanced benefits of antibody
effector functions.
362
6 miRNA profiling of the natural antibody producing B-cell
It has already been suggested that lessons from nature can be explored in the
identification of attractive engineering targets for CHO cell manipulation, B-cells
being one. Differential miRNA expression profiling could be explored on resting
B-cells and activated B-cells with subsequent screening of these identified
miRNA candidates in CHO.
363
Bibliography
Addya, S., Mullick, J., Fang, J.K. & Avadhani, N.G. 1994, "Purification and
characterization of a hepatic mitochondrial glutathione S-transferase exhibiting
immunochemical relationship to the alpha-class of cytosolic isoenzymes", Archives
of Biochemistry and Biophysics, vol. 310, no. 1, pp. 82-88.
Aitken, A. 2006, "14-3-3 Proteins: a Historic Overview", Seminars in cancer biology,
vol. 16, no. 3, pp. 162-172.
Alexiou, P., Maragkakis, M., Papadopoulos, G.L., Reczko, M. & Hatzigeorgiou, A.G.
2009, "Lost in translation: an assessment and perspective for computational
microRNA target identification", Bioinformatics (Oxford, England), vol. 25, no. 23,
pp. 3049-3055.
Al-Fageeh, M.B., Marchant, R.J., Carden, M.J. & Smales, C.M. 2006, "The cold-shock
response in cultured mammalian cells: harnessing the response for the improvement
of recombinant protein production", Biotechnology and bioengineering, vol. 93, no.
5, pp. 829-835.
Altamirano, C., Paredes, C., Cairo, J.J. & Godia, F. 2000, "Improvement of CHO cell
culture medium formulation: simultaneous substitution of glucose and glutamine",
Biotechnology progress, vol. 16, no. 1, pp. 69-75.
Altamirano, C., Paredes, C., Illanes, A., Cairo, J.J. & Godia, F. 2004, "Strategies for fed-
batch cultivation of t-PA producing CHO cells: substitution of glucose and glutamine
and rational design of culture medium", Journal of Biotechnology, vol. 110, no. 2,
pp. 171-179.
Ambros, V. 2008, "The evolution of our thinking about microRNAs", Nature medicine,
vol. 14, no. 10, pp. 1036-1040.
Ambros, V., Bartel, B., Bartel, D.P., Burge, C.B., Carrington, J.C., Chen, X., Dreyfuss,
G., Eddy, S.R., Griffiths-Jones, S., Marshall, M., Matzke, M., Ruvkun, G. & Tuschl,
T. 2003, "A uniform system for microRNA annotation", RNA (New York, N.Y.), vol.
9, no. 3, pp. 277-279.
An, J., Pan, Y., Yan, Z., Li, W., Cui, J., Jiao, Y., Tian, L., Xing, R. & Lu, Y. 2013, "MiR-
23a in Amplified 19p13.13 Loci Targets Metallothionein 2A and Promotes Growth
in Gastric Cancer Cells", Journal of cellular biochemistry, .
Arden, N. & Betenbaugh, M.J. 2004, "Life and death in mammalian cell culture:
strategies for apoptosis inhibition", Trends in biotechnology, vol. 22, no. 4, pp. 174-
180.
Arvey, A., Larsson, E., Sander, C., Leslie, C.S. & Marks, D.S. 2010, "Target mRNA
abundance dilutes microRNA and siRNA activity", Molecular systems biology, vol.
6, pp. 363.
364
Bader, A.G. 2012, "miR-34 - a microRNA replacement therapy is headed to the clinic",
Frontiers in genetics, vol. 3, pp. 120.
Bagga, S., Bracht, J., Hunter, S., Massirer, K., Holtz, J., Eachus, R. & Pasquinelli, A.E.
2005, "Regulation by let-7 and lin-4 miRNAs results in target mRNA degradation",
Cell, vol. 122, no. 4, pp. 553-563.
Baik, J.Y., Lee, M.S., An, S.R., Yoon, S.K., Joo, E.J., Kim, Y.H., Park, H.W. & Lee,
G.M. 2006, "Initial transcriptome and proteome analyses of low culture temperature-
induced expression in CHO cells producing erythropoietin", Biotechnology and
bioengineering, vol. 93, no. 2, pp. 361-371.
Bail, S., Swerdel, M., Liu, H., Jiao, X., Goff, L.A., Hart, R.P. & Kiledjian, M. 2010,
"Differential regulation of microRNA stability", RNA (New York, N.Y.), vol. 16, no.
5, pp. 1032-1039.
Barron, N., Keenan, J., Gammell, P., Martinez, V.G., Freeman, A., Masters, J.R. &
Clynes, M. 2012, "Biochemical relapse following radical prostatectomy and miR-
200a levels in prostate cancer", The Prostate, vol. 72, no. 11, pp. 1193-1199.
Barron, N., Kumar, N., Sanchez, N., Doolan, P., Clarke, C., Meleady, P., O'Sullivan, F.
& Clynes, M. 2011a, "Engineering CHO cell growth and recombinant protein
productivity by overexpression of miR-7", Journal of Biotechnology, vol. 151, no.
2, pp. 204-211.
Barron, N., Kumar, N., Sanchez, N., Doolan, P., Clarke, C., Meleady, P., O'Sullivan, F.
& Clynes, M. 2011b, "Engineering CHO cell growth and recombinant protein
productivity by overexpression of miR-7", Journal of Biotechnology, vol. 151, no.
2, pp. 204-211.
Barron, N., Kumar, N., Sanchez, N., Doolan, P., Clarke, C., Meleady, P., O'Sullivan, F.
& Clynes, M. 2011c, "Engineering CHO cell growth and recombinant protein
productivity by overexpression of miR-7", Journal of Biotechnology, vol. 151, no.
2, pp. 204-211.
Barron, N., Sanchez, N., Kelly, P. & Clynes, M. 2011d, "MicroRNAs: tiny targets for
engineering CHO cell phenotypes?", Biotechnology Letters, vol. 33, no. 1, pp. 11-
21.
Barron, N., Sanchez, N., Kelly, P. & Clynes, M. 2011e, "MicroRNAs: tiny targets for
engineering CHO cell phenotypes?", Biotechnology Letters, vol. 33, no. 1, pp. 11-
21.
Bartel, D.P. 2009, "MicroRNAs: target recognition and regulatory functions", Cell, vol.
136, no. 2, pp. 215-233.
Bartel, D.P. 2004, "MicroRNAs: Genomics, Biogenesis, Mechanism, and Function",
Cell, vol. 116, no. 2, pp. 281-297.
365
Bartlett, D.W. & Davis, M.E. 2006, "Insights into the kinetics of siRNA-mediated gene
silencing from live-cell and live-animal bioluminescent imaging", Nucleic acids
research, vol. 34, no. 1, pp. 322-333.
Bauerschmitt, H., Mick, D.U., Deckers, M., Vollmer, C., Funes, S., Kehrein, K., Ott, M.,
Rehling, P. & Herrmann, J.M. 2010, "Ribosome-binding proteins Mdm38 and Mba1
display overlapping functions for regulation of mitochondrial translation",
Molecular biology of the cell, vol. 21, no. 12, pp. 1937-1944.
Baumgart, E., Fahimi, H.D., Stich, A. & Volkl, A. 1996, "L-lactate dehydrogenase A4-
and A3B isoforms are bona fide peroxisomal enzymes in rat liver. Evidence for
involvement in intraperoxisomal NADH reoxidation", The Journal of biological
chemistry, vol. 271, no. 7, pp. 3846-3855.
Becker, E., Florin, L., Pfizenmaier, K. & Kaufmann, H. 2008, "An XBP-1 dependent
bottle-neck in production of IgG subtype antibodies in chemically defined serum-
free Chinese hamster ovary (CHO) fed-batch processes", Journal of Biotechnology,
vol. 135, no. 2, pp. 217-223.
Becker, J., Hackl, M., Rupp, O., Jakobi, T., Schneider, J., Szczepanowski, R., Bekel, T.,
Borth, N., Goesmann, A., Grillari, J., Kaltschmidt, C., Noll, T., Puhler, A., Tauch,
A. & Brinkrolf, K. 2011, "Unraveling the Chinese hamster ovary cell line
transcriptome by next-generation sequencing", Journal of Biotechnology, vol. 156,
no. 3, pp. 227-235.
Bensinger, S.J. & Christofk, H.R. 2012, "New aspects of the Warburg effect in cancer
cell biology", Seminars in cell & developmental biology, vol. 23, no. 4, pp. 352-361.
Bernardi, C., Soffientini, U., Piacente, F. & Tonetti, M.G. 2013a, "Effects of MicroRNAs
on Fucosyltransferase 8 (FUT8) Expression in Hepatocarcinoma Cells", PloS one,
vol. 8, no. 10, pp. e76540.
Bernardi, C., Soffientini, U., Piacente, F. & Tonetti, M.G. 2013b, "Effects of microRNAs
on fucosyltransferase 8 (FUT8) expression in hepatocarcinoma cells", PloS one, vol.
8, no. 10, pp. e76540.
Bernstein, E., Kim, S.Y., Carmell, M.A., Murchison, E.P., Alcorn, H., Li, M.Z., Mills,
A.A., Elledge, S.J., Anderson, K.V. & Hannon, G.J. 2003, "Dicer is essential for
mouse development", Nature genetics, vol. 35, no. 3, pp. 215-217.
Beuger, V., Kunkele, K.P., Koll, H., Gartner, A., Bahner, M., Burtscher, H. & Klein, C.
2009, "Short-hairpin-RNA-mediated silencing of fucosyltransferase 8 in Chinese-
hamster ovary cells for the production of antibodies with enhanced antibody immune
effector function", Biotechnology and applied biochemistry, vol. 53, no. Pt 1, pp. 31-
37.
Bhagwat, S.V., Vijayasarathy, C., Raza, H., Mullick, J. & Avadhani, N.G. 1998,
"Preferential effects of nicotine and 4-(N-methyl-N-nitrosamine)-1-(3-pyridyl)-1-
butanone on mitochondrial glutathione S-transferase A4-4 induction and increased
366
oxidative stress in the rat brain", Biochemical pharmacology, vol. 56, no. 7, pp. 831-
839.
Bi, J.X., Shuttleworth, J. & Al-Rubeai, M. 2004, "Uncoupling of cell growth and
proliferation results in enhancement of productivity in p21CIP1-arrested CHO cells",
Biotechnology and bioengineering, vol. 85, no. 7, pp. 741-749.
Bian, H.B., Pan, X., Yang, J.S., Wang, Z.X. & De, W. 2011, "Upregulation of microRNA-
451 increases cisplatin sensitivity of non-small cell lung cancer cell line (A549)",
Journal of experimental & clinical cancer research : CR, vol. 30, pp. 20-9966-30-
20.
Bindu, S., Pal, C., Dey, S., Goyal, M., Alam, A., Iqbal, M.S., Dutta, S., Sarkar, S., Kumar,
R., Maity, P. & Bandyopadhyay, U. 2011, "Translocation of heme oxygenase-1 to
mitochondria is a novel cytoprotective mechanism against non-steroidal anti-
inflammatory drug-induced mitochondrial oxidative stress, apoptosis, and gastric
mucosal injury", The Journal of biological chemistry, vol. 286, no. 45, pp. 39387-
39402.
Birch, J.R. & Arathoon, R. 1990, "Suspension culture of mammalian cells", Bioprocess
technology, vol. 10, pp. 251-270.
Bleckwenn, N.A. & Shiloach, J. 2004, "Large-scale cell culture", Current protocols in
immunology / edited by John E.Coligan et al., vol. Appendix 1, pp. Appendix 1U.
Bohnsack, M.T., Czaplinski, K. & Gorlich, D. 2004, "Exportin 5 is a RanGTP-dependent
dsRNA-binding protein that mediates nuclear export of pre-miRNAs", RNA (New
York, N.Y.), vol. 10, no. 2, pp. 185-191.
Boland, A., Huntzinger, E., Schmidt, S., Izaurralde, E. & Weichenrieder, O. 2011,
"Crystal structure of the MID-PIWI lobe of a eukaryotic Argonaute protein",
Proceedings of the National Academy of Sciences of the United States of America,
vol. 108, no. 26, pp. 10466-10471.
Bommer, G.T., Gerin, I., Feng, Y., Kaczorowski, A.J., Kuick, R., Love, R.E., Zhai, Y.,
Giordano, T.J., Qin, Z.S., Moore, B.B., MacDougald, O.A., Cho, K.R. & Fearon,
E.R. 2007, "p53-mediated activation of miRNA34 candidate tumor-suppressor
genes", Current biology : CB, vol. 17, no. 15, pp. 1298-1307.
Borth, N., Mattanovich, D., Kunert, R. & Katinger, H. 2005, "Effect of increased
expression of protein disulfide isomerase and heavy chain binding protein on
antibody secretion in a recombinant CHO cell line", Biotechnology progress, vol.
21, no. 1, pp. 106-111.
Bortolin-Cavaille, M.L., Dance, M., Weber, M. & Cavaille, J. 2009, "C19MC
microRNAs are processed from introns of large Pol-II, non-protein-coding
transcripts", Nucleic acids research, vol. 37, no. 10, pp. 3464-3473.
367
Bradley, S.A., Ouyang, A., Purdie, J., Smitka, T.A., Wang, T. & Kaerner, A. 2010,
"Fermentanomics: monitoring mammalian cell cultures with NMR spectroscopy",
Journal of the American Chemical Society, vol. 132, no. 28, pp. 9531-9533.
Bravo, R., Vicencio, J.M., Parra, V., Troncoso, R., Munoz, J.P., Bui, M., Quiroga, C.,
Rodriguez, A.E., Verdejo, H.E., Ferreira, J., Iglewski, M., Chiong, M., Simmen, T.,
Zorzano, A., Hill, J.A., Rothermel, B.A., Szabadkai, G. & Lavandero, S. 2011,
"Increased ER-mitochondrial coupling promotes mitochondrial respiration and
bioenergetics during early phases of ER stress", Journal of cell science, vol. 124, no.
Pt 13, pp. 2143-2152.
Brennecke, J., Stark, A., Russell, R.B. & Cohen, S.M. 2005, "Principles of microRNA-
target recognition", PLoS biology, vol. 3, no. 3, pp. e85.
Brooks, A.R., Harkins, R.N., Wang, P., Qian, H.S., Liu, P. & Rubanyi, G.M. 2004,
"Transcriptional silencing is associated with extensive methylation of the CMV
promoter following adenoviral gene delivery to muscle", The journal of gene
medicine, vol. 6, no. 4, pp. 395-404.
Bruggemann, M., Williams, G.T., Bindon, C.I., Clark, M.R., Walker, M.R., Jefferis, R.,
Waldmann, H. & Neuberger, M.S. 1987, "Comparison of the effector functions of
human immunoglobulins using a matched set of chimeric antibodies", The Journal
of experimental medicine, vol. 166, no. 5, pp. 1351-1361.
Bruhns, P., Iannascoli, B., England, P., Mancardi, D.A., Fernandez, N., Jorieux, S. &
Daeron, M. 2009, "Specificity and affinity of human Fcgamma receptors and their
polymorphic variants for human IgG subclasses", Blood, vol. 113, no. 16, pp. 3716-
3725.
Bu, Y., Lu, C., Bian, C., Wang, J., Li, J., Zhang, B., Li, Z., Brewer, G. & Zhao, R.C.
2009, "Knockdown of Dicer in MCF-7 human breast carcinoma cells results in G1
arrest and increased sensitivity to cisplatin", Oncology reports, vol. 21, no. 1, pp. 13-
17.
Burd, C.E., Jeck, W.R., Liu, Y., Sanoff, H.K., Wang, Z. & Sharpless, N.E. 2010,
"Expression of linear and novel circular forms of an INK4/ARF-associated non-
coding RNA correlates with atherosclerosis risk", PLoS genetics, vol. 6, no. 12, pp.
e1001233.
Butler, M. 2005, "Animal cell cultures: recent achievements and perspectives in the
production of biopharmaceuticals", Applied Microbiology and Biotechnology, vol.
68, no. 3, pp. 283-291.
Butler, M. & Meneses-Acosta, A. 2012, "Recent advances in technology supporting
biopharmaceutical production from mammalian cells", Applied Microbiology and
Biotechnology, vol. 96, no. 4, pp. 885-894.
Byrd, A.E., Aragon, I.V. & Brewer, J.W. 2012, "MicroRNA-30c-2* limits expression of
proadaptive factor XBP1 in the unfolded protein response", The Journal of cell
biology, vol. 196, no. 6, pp. 689-698.
368
Cacciatore, J.J., Chasin, L.A. & Leonard, E.F. 2010, "Gene amplification and vector
engineering to achieve rapid and high-level therapeutic protein production using the
Dhfr-based CHO cell selection system", Biotechnology Advances, vol. 28, no. 6, pp.
673-681.
Cadenas, C., Franckenstein, D., Schmidt, M., Gehrmann, M., Hermes, M., Geppert, B.,
Schormann, W., Maccoux, L.J., Schug, M., Schumann, A., Wilhelm, C., Freis, E.,
Ickstadt, K., Rahnenfuhrer, J., Baumbach, J.I., Sickmann, A. & Hengstler, J.G. 2010,
"Role of thioredoxin reductase 1 and thioredoxin interacting protein in prognosis of
breast cancer", Breast cancer research : BCR, vol. 12, no. 3, pp. R44.
Cai, H., Wang, C.C. & Tsou, C.L. 1994, "Chaperone-like activity of protein disulfide
isomerase in the refolding of a protein with no disulfide bonds", The Journal of
biological chemistry, vol. 269, no. 40, pp. 24550-24552.
Cai, X., Hagedorn, C.H. & Cullen, B.R. 2004, "Human microRNAs are processed from
capped, polyadenylated transcripts that can also function as mRNAs", RNA (New
York, N.Y.), vol. 10, no. 12, pp. 1957-1966.
Calin, G.A., Cimmino, A., Fabbri, M., Ferracin, M., Wojcik, S.E., Shimizu, M., Taccioli,
C., Zanesi, N., Garzon, R., Aqeilan, R.I., Alder, H., Volinia, S., Rassenti, L., Liu, X.,
Liu, C.G., Kipps, T.J., Negrini, M. & Croce, C.M. 2008, "MiR-15a and miR-16-1
cluster functions in human leukemia", Proceedings of the National Academy of
Sciences of the United States of America, vol. 105, no. 13, pp. 5166-5171.
Campbell, C.T., Kolesar, J.E. & Kaufman, B.A. 2012, "Mitochondrial transcription factor
A regulates mitochondrial transcription initiation, DNA packaging, and genome
copy number", Biochimica et biophysica acta, vol. 1819, no. 9-10, pp. 921-929.
Cao, Z., Zhang, R., Li, J., Huang, H., Zhang, D., Zhang, J., Gao, J., Chen, J. & Huang, C.
2013, "X-linked inhibitor of apoptosis protein (XIAP) regulation of cyclin D1 protein
expression and cancer cell anchorage-independent growth via its E3 ligase-mediated
protein phosphatase 2A/c-Jun axis", The Journal of biological chemistry, vol. 288,
no. 28, pp. 20238-20247.
Carlage, T., Hincapie, M., Zang, L., Lyubarskaya, Y., Madden, H., Mhatre, R. &
Hancock, W.S. 2009a, "Proteomic profiling of a high-producing Chinese hamster
ovary cell culture", Analytical Chemistry, vol. 81, no. 17, pp. 7357-7362.
Carlage, T., Hincapie, M., Zang, L., Lyubarskaya, Y., Madden, H., Mhatre, R. &
Hancock, W.S. 2009b, "Proteomic profiling of a high-producing Chinese hamster
ovary cell culture", Analytical Chemistry, vol. 81, no. 17, pp. 7357-7362.
Carrington, J.C. & Ambros, V. 2003, "Role of MicroRNAs in Plant and Animal
Development", Science, vol. 301, no. 5631, pp. 336-338.
Carroll, J., Fearnley, I.M., Skehel, J.M., Shannon, R.J., Hirst, J. & Walker, J.E. 2006,
"Bovine complex I is a complex of 45 different subunits", The Journal of biological
chemistry, vol. 281, no. 43, pp. 32724-32727.
369
Carvalhal, A.V., Marcelino, I. & Carrondo, M.J. 2003, "Metabolic changes during cell
growth inhibition by p27 overexpression", Applied Microbiology and Biotechnology,
vol. 63, no. 2, pp. 164-173.
Casagrande, R., Stern, P., Diehn, M., Shamu, C., Osario, M., Zuniga, M., Brown, P.O. &
Ploegh, H. 2000, "Degradation of proteins from the ER of S. cerevisiae requires an
intact unfolded protein response pathway", Molecular cell, vol. 5, no. 4, pp. 729-735.
Cenas, N., Nivinskas, H., Anusevicius, Z., Sarlauskas, J., Lederer, F. & Arner, E.S. 2004,
"Interactions of quinones with thioredoxin reductase: a challenge to the antioxidant
role of the mammalian selenoprotein", The Journal of biological chemistry, vol. 279,
no. 4, pp. 2583-2592.
Chan, A.C. & Carter, P.J. 2010, "Therapeutic antibodies for autoimmunity and
inflammation", Nature reviews.Immunology, vol. 10, no. 5, pp. 301-316.
Chan, J.A., Krichevsky, A.M. & Kosik, K.S. 2005, "MicroRNA-21 is an antiapoptotic
factor in human glioblastoma cells", Cancer research, vol. 65, no. 14, pp. 6029-6033.
Chang, C., Xu, K. & Shu, H. 2011, "The role of isocitrate dehydrogenase mutations in
glioma brain tumors", Molecular targets of CNS tumors, pp. 413-436.
Chang, K.W., Kao, S.Y., Wu, Y.H., Tsai, M.M., Tu, H.F., Liu, C.J., Lui, M.T. & Lin,
S.C. 2013, "Passenger strand miRNA miR-31* regulates the phenotypes of oral
cancer cells by targeting RhoA", Oral oncology, vol. 49, no. 1, pp. 27-33.
Chang, T.C., Wentzel, E.A., Kent, O.A., Ramachandran, K., Mullendore, M., Lee, K.H.,
Feldmann, G., Yamakuchi, M., Ferlito, M., Lowenstein, C.J., Arking, D.E., Beer,
M.A., Maitra, A. & Mendell, J.T. 2007, "Transactivation of miR-34a by p53 broadly
influences gene expression and promotes apoptosis", Molecular cell, vol. 26, no. 5,
pp. 745-752.
Chatterjee, S., Fasler, M., Bussing, I. & Grosshans, H. 2011, "Target-mediated protection
of endogenous microRNAs in C. elegans", Developmental cell, vol. 20, no. 3, pp.
388-396.
Chatterjee, S. & Grosshans, H. 2009, "Active turnover modulates mature microRNA
activity in Caenorhabditis elegans", Nature, vol. 461, no. 7263, pp. 546-549.
Chen, C.Z., Li, L., Lodish, H.F. & Bartel, D.P. 2004, "MicroRNAs modulate
hematopoietic lineage differentiation", Science (New York, N.Y.), vol. 303, no. 5654,
pp. 83-86.
Chen, F., Ye, Z., Zhao, L., Liu, X., Fan, L. & Tan, W.S. 2012, "Correlation of antibody
production rate with glucose and lactate metabolism in Chinese hamster ovary cells",
Biotechnology Letters, vol. 34, no. 3, pp. 425-432.
Chen, K., Liu, Q., Xie, L., Sharp, P.A. & Wang, D.I. 2001, "Engineering of a mammalian
cell line for reduction of lactate formation and high monoclonal antibody
production", Biotechnology and bioengineering, vol. 72, no. 1, pp. 55-61.
370
Chen, Q., Xu, J., Li, L., Li, H., Mao, S., Zhang, F., Zen, K., Zhang, C.Y. & Zhang, Q.
2014, "MicroRNA-23a/b and microRNA-27a/b suppress Apaf-1 protein and
alleviate hypoxia-induced neuronal apoptosis", Cell death & disease, vol. 5, pp.
e1132.
Chen, Q.R., Yu, L.R., Tsang, P., Wei, J.S., Song, Y.K., Cheuk, A., Chung, J.Y., Hewitt,
S.M., Veenstra, T.D. & Khan, J. 2011a, "Systematic proteome analysis identifies
transcription factor YY1 as a direct target of miR-34a", Journal of proteome
research, vol. 10, no. 2, pp. 479-487.
Chen, S., Synowsky, S., Tinti, M. & MacKintosh, C. 2011b, "The capture of
phosphoproteins by 14-3-3 proteins mediates actions of insulin", Trends in
endocrinology and metabolism: TEM, vol. 22, no. 11, pp. 429-436.
Chen, Z., Lu, W., Garcia-Prieto, C. & Huang, P. 2007, "The Warburg effect and its cancer
therapeutic implications", Journal of Bioenergetics and Biomembranes, vol. 39, no.
3, pp. 267-274.
Chendrimada, T.P., Gregory, R.I., Kumaraswamy, E., Norman, J., Cooch, N., Nishikura,
K. & Shiekhattar, R. 2005, "TRBP recruits the Dicer complex to Ago2 for
microRNA processing and gene silencing", Nature, vol. 436, no. 7051, pp. 740-744.
Chhabra, R., Adlakha, Y.K., Hariharan, M., Scaria, V. & Saini, N. 2009, "Upregulation
of miR-23a-27a-24-2 cluster induces caspase-dependent and -independent apoptosis
in human embryonic kidney cells", PloS one, vol. 4, no. 6, pp. e5848.
Chhabra, R., Dubey, R. & Saini, N. 2010a, "Cooperative and individualistic functions of
the microRNAs in the miR-23a~27a~24-2 cluster and its implication in human
diseases", Molecular cancer, vol. 9, pp. 232-4598-9-232.
Chhabra, R., Dubey, R. & Saini, N. 2010b, "Cooperative and individualistic functions of
the microRNAs in the miR-23a~27a~24-2 cluster and its implication in human
diseases", Molecular cancer, vol. 9, pp. 232-4598-9-232.
Chi, S.W., Hannon, G.J. & Darnell, R.B. 2012, "An alternative mode of microRNA target
recognition", Nature structural & molecular biology, vol. 19, no. 3, pp. 321-327.
Chi, S.W., Zang, J.B., Mele, A. & Darnell, R.B. 2009, "Argonaute HITS-CLIP decodes
microRNA-mRNA interaction maps", Nature, vol. 460, no. 7254, pp. 479-486.
Chiang, G.G. & Sisk, W.P. 2005, "Bcl-x(L) mediates increased production of humanized
monoclonal antibodies in Chinese hamster ovary cells", Biotechnology and
bioengineering, vol. 91, no. 7, pp. 779-792.
Chiang, H.R., Schoenfeld, L.W., Ruby, J.G., Auyeung, V.C., Spies, N., Baek, D.,
Johnston, W.K., Russ, C., Luo, S., Babiarz, J.E., Blelloch, R., Schroth, G.P.,
Nusbaum, C. & Bartel, D.P. 2010, "Mammalian microRNAs: experimental
evaluation of novel and previously annotated genes", Genes & development, vol. 24,
no. 10, pp. 992-1009.
371
Choi, Y.J., Lin, C.P., Ho, J.J., He, X., Okada, N., Bu, P., Zhong, Y., Kim, S.Y., Bennett,
M.J., Chen, C., Ozturk, A., Hicks, G.G., Hannon, G.J. & He, L. 2011, "miR-34
miRNAs provide a barrier for somatic cell reprogramming", Nature cell biology, vol.
13, no. 11, pp. 1353-1360.
Chong, W.P., Goh, L.T., Reddy, S.G., Yusufi, F.N., Lee, D.Y., Wong, N.S., Heng, C.K.,
Yap, M.G. & Ho, Y.S. 2009a, "Metabolomics profiling of extracellular metabolites
in recombinant Chinese Hamster Ovary fed-batch culture", Rapid communications
in mass spectrometry : RCM, vol. 23, no. 23, pp. 3763-3771.
Chong, W.P., Goh, L.T., Reddy, S.G., Yusufi, F.N., Lee, D.Y., Wong, N.S., Heng, C.K.,
Yap, M.G. & Ho, Y.S. 2009b, "Metabolomics profiling of extracellular metabolites
in recombinant Chinese Hamster Ovary fed-batch culture", Rapid communications
in mass spectrometry : RCM, vol. 23, no. 23, pp. 3763-3771.
Chong, W.P., Reddy, S.G., Yusufi, F.N., Lee, D.Y., Wong, N.S., Heng, C.K., Yap, M.G.
& Ho, Y.S. 2010, "Metabolomics-driven approach for the improvement of Chinese
hamster ovary cell growth: overexpression of malate dehydrogenase II", Journal of
Biotechnology, vol. 147, no. 2, pp. 116-121.
Chong, W.P., Thng, S.H., Hiu, A.P., Lee, D.Y., Chan, E.C. & Ho, Y.S. 2012, "LC-MS-
based metabolic characterization of high monoclonal antibody-producing Chinese
hamster ovary cells", Biotechnology and bioengineering, vol. 109, no. 12, pp. 3103-
3111.
Chong, W.P., Yusufi, F.N., Lee, D.Y., Reddy, S.G., Wong, N.S., Heng, C.K., Yap, M.G.
& Ho, Y.S. 2011, "Metabolomics-based identification of apoptosis-inducing
metabolites in recombinant fed-batch CHO culture media", Journal of
Biotechnology, vol. 151, no. 2, pp. 218-224.
Chotteau, V. & Lindqvist, A. 2012, "Study of the Effect of High pH and Alkali Addition
in a Cultivation of Chinese Hamster Ovary Cell", Proceedings of the 21st Annual
Meeting of the European Society for Animal Cell Technology (ESACT), Dublin,
Ireland, June 7-10, 2009 Springer, pp. 323.
Christoffersen, N.R., Shalgi, R., Frankel, L.B., Leucci, E., Lees, M., Klausen, M., Pilpel,
Y., Nielsen, F.C., Oren, M. & Lund, A.H. 2010, "p53-independent upregulation of
miR-34a during oncogene-induced senescence represses MYC", Cell death and
differentiation, vol. 17, no. 2, pp. 236-245.
Chung, J.Y., Lim, S.W., Hong, Y.J., Hwang, S.O. & Lee, G.M. 2004, "Effect of
doxycycline-regulated calnexin and calreticulin expression on specific
thrombopoietin productivity of recombinant Chinese hamster ovary cells",
Biotechnology and bioengineering, vol. 85, no. 5, pp. 539-546.
Clarke, C., Doolan, P., Barron, N., Meleady, P., O'Sullivan, F., Gammell, P., Melville,
M., Leonard, M. & Clynes, M. 2011a, "Large scale microarray profiling and
coexpression network analysis of CHO cells identifies transcriptional modules
associated with growth and productivity", Journal of Biotechnology,20:155(3), pp.
530-9.
372
Clarke, C., Doolan, P., Barron, N., Meleady, P., O'Sullivan, F., Gammell, P., Melville,
M., Leonard, M. & Clynes, M. 2011b, "Predicting cell-specific productivity from
CHO gene expression", Journal of Biotechnology, vol. 151, no. 2, pp. 159-165.
Clarke, C., Henry, M., Doolan, P., Kelly, S., Aherne, S., Sanchez, N., Kelly, P., Kinsella,
P., Breen, L., Madden, S.F., Zhang, L., Leonard, M., Clynes, M., Meleady, P. &
Barron, N. 2012, "Integrated miRNA, mRNA and protein expression analysis reveals
the role of post-transcriptional regulation in controlling CHO cell growth rate", BMC
genomics, vol. 13, pp. 656-2164-13-656.
Clincke, M.F., Molleryd, C., Zhang, Y., Lindskog, E., Walsh, K. & Chotteau, V. 2013,
"Very high density of CHO cells in perfusion by ATF or TFF in WAVE bioreactor.
Part I. Effect of the cell density on the process", Biotechnology progress, vol. 29, no.
3, pp. 754-767.
Cole, K.A., Attiyeh, E.F., Mosse, Y.P., Laquaglia, M.J., Diskin, S.J., Brodeur, G.M. &
Maris, J.M. 2008, "A functional screen identifies miR-34a as a candidate
neuroblastoma tumor suppressor gene", Molecular cancer research : MCR, vol. 6,
no. 5, pp. 735-742.
Conaway, R.C. & Conaway, J.W. 1988, "ATP activates transcription initiation from
promoters by RNA polymerase II in a reversible step prior to RNA synthesis", The
Journal of biological chemistry, vol. 263, no. 6, pp. 2962-2968.
Congy-Jolivet, N., Probst, A., Watier, H. & Thibault, G. 2007, "Recombinant therapeutic
monoclonal antibodies: mechanisms of action in relation to structural and functional
duality", Critical reviews in oncology/hematology, vol. 64, no. 3, pp. 226-233.
Corney, D.C., Hwang, C.I., Matoso, A., Vogt, M., Flesken-Nikitin, A., Godwin, A.K.,
Kamat, A.A., Sood, A.K., Ellenson, L.H., Hermeking, H. & Nikitin, A.Y. 2010,
"Frequent downregulation of miR-34 family in human ovarian cancers", Clinical
cancer research : an official journal of the American Association for Cancer
Research, vol. 16, no. 4, pp. 1119-1128.
Cost, G.J., Freyvert, Y., Vafiadis, A., Santiago, Y., Miller, J.C., Rebar, E., Collingwood,
T.N., Snowden, A. & Gregory, P.D. 2010, "BAK and BAX deletion using zinc-finger
nucleases yields apoptosis-resistant CHO cells", Biotechnology and bioengineering,
vol. 105, no. 2, pp. 330-340.
Danan, M., Schwartz, S., Edelheit, S. & Sorek, R. 2012, "Transcriptome-wide discovery
of circular RNAs in Archaea", Nucleic acids research, vol. 40, no. 7, pp. 3131-3142.
Dang, C.V. 2010, "Rethinking the Warburg effect with Myc micromanaging glutamine
metabolism", Cancer research, vol. 70, no. 3, pp. 859-862.
das Neves, R.P., Jones, N.S., Andreu, L., Gupta, R., Enver, T. & Iborra, F.J. 2010,
"Connecting variability in global transcription rate to mitochondrial variability",
PLoS biology, vol. 8, no. 12, pp. e1000560.
373
Datta, P., Linhardt, R.J. & Sharfstein, S.T. 2013, "An 'omics approach towards CHO cell
engineering", Biotechnology and bioengineering, vol. 110, no. 5, pp. 1255-1271.
Davalos, A., Goedeke, L., Smibert, P., Ramirez, C.M., Warrier, N.P., Andreo, U., Cirera-
Salinas, D., Rayner, K., Suresh, U., Pastor-Pareja, J.C., Esplugues, E., Fisher, E.A.,
Penalva, L.O., Moore, K.J., Suarez, Y., Lai, E.C. & Fernandez-Hernando, C. 2011,
"miR-33a/b contribute to the regulation of fatty acid metabolism and insulin
signaling", Proceedings of the National Academy of Sciences of the United States of
America, vol. 108, no. 22, pp. 9232-9237.
Davie, J.R. 2003, "Inhibition of histone deacetylase activity by butyrate", The Journal of
nutrition, vol. 133, no. 7 Suppl, pp. 2485S-2493S.
Davies, J., Jiang, L., Pan, L.Z., LaBarre, M.J., Anderson, D. & Reff, M. 2001,
"Expression of GnTIII in a recombinant anti-CD20 CHO production cell line:
Expression of antibodies with altered glycoforms leads to an increase in ADCC
through higher affinity for FC gamma RIII", Biotechnology and bioengineering, vol.
74, no. 4, pp. 288-294.
Davis, R., Schooley, K., Rasmussen, B., Thomas, J. & Reddy, P. 2000, "Effect of PDI
overexpression on recombinant protein secretion in CHO cells", Biotechnology
progress, vol. 16, no. 5, pp. 736-743.
De Jesus, M. & Wurm, F.M. 2011, "Manufacturing recombinant proteins in kg-ton
quantities using animal cells in bioreactors", European journal of pharmaceutics and
biopharmaceutics : official journal of Arbeitsgemeinschaft fur Pharmazeutische
Verfahrenstechnik e.V, vol. 78, no. 2, pp. 184-188.
De Leon Gatti, M., Wlaschin, K.F., Nissom, P.M., Yap, M. & Hu, W.S. 2007,
"Comparative transcriptional analysis of mouse hybridoma and recombinant Chinese
hamster ovary cells undergoing butyrate treatment", Journal of bioscience and
bioengineering, vol. 103, no. 1, pp. 82-91.
Dean, N.M. 2001, "Functional genomics and target validation approaches using antisense
oligonucleotide technology", Current opinion in biotechnology, vol. 12, no. 6, pp.
622-625.
DeBerardinis, R.J., Mancuso, A., Daikhin, E., Nissim, I., Yudkoff, M., Wehrli, S. &
Thompson, C.B. 2007, "Beyond aerobic glycolysis: transformed cells can engage in
glutamine metabolism that exceeds the requirement for protein and nucleotide
synthesis", Proceedings of the National Academy of Sciences of the United States of
America, vol. 104, no. 49, pp. 19345-19350.
Dechant, M., Beyer, T., Schneider-Merck, T., Weisner, W., Peipp, M., van de Winkel,
J.G. & Valerius, T. 2007, "Effector mechanisms of recombinant IgA antibodies
against epidermal growth factor receptor", Journal of immunology (Baltimore, Md.:
1950), vol. 179, no. 5, pp. 2936-2943.
Dell'Antone, P. 2012, "Energy metabolism in cancer cells: how to explain the Warburg
and Crabtree effects?", Medical hypotheses, vol. 79, no. 3, pp. 388-392.
374
DeMaria, C.T., Cairns, V., Schwarz, C., Zhang, J., Guerin, M., Zuena, E., Estes, S. &
Karey, K.P. 2008, "Accelerated Clone Selection for Recombinant CHO Cells Using
a FACS‐Based High‐Throughput Screen", Biotechnology progress, vol. 23, no. 2,
pp. 465-472.
Dews, M., Homayouni, A., Yu, D., Murphy, D., Sevignani, C., Wentzel, E., Furth, E.E.,
Lee, W.M., Enders, G.H., Mendell, J.T. & Thomas-Tikhonenko, A. 2006a,
"Augmentation of tumor angiogenesis by a Myc-activated microRNA cluster",
Nature genetics, vol. 38, no. 9, pp. 1060-1065.
Dews, M., Homayouni, A., Yu, D., Murphy, D., Sevignani, C., Wentzel, E., Furth, E.E.,
Lee, W.M., Enders, G.H., Mendell, J.T. & Thomas-Tikhonenko, A. 2006b,
"Augmentation of tumor angiogenesis by a Myc-activated microRNA cluster",
Nature genetics, vol. 38, no. 9, pp. 1060-1065.
Dheda, K., Huggett, J.F., Bustin, S.A., Johnson, M.A., Rook, G. & Zumla, A. 2004,
"Validation of housekeeping genes for normalizing RNA expression in real-time
PCR", BioTechniques, vol. 37, no. 1, pp. 112-4, 116, 118-9.
Diehn, M., Cho, R.W., Lobo, N.A., Kalisky, T., Dorie, M.J., Kulp, A.N., Qian, D., Lam,
J.S., Ailles, L.E., Wong, M., Joshua, B., Kaplan, M.J., Wapnir, I., Dirbas, F.M.,
Somlo, G., Garberoglio, C., Paz, B., Shen, J., Lau, S.K., Quake, S.R., Brown, J.M.,
Weissman, I.L. & Clarke, M.F. 2009, "Association of reactive oxygen species levels
and radioresistance in cancer stem cells", Nature, vol. 458, no. 7239, pp. 780-783.
Dill, H., Linder, B., Fehr, A. & Fischer, U. 2012, "Intronic miR-26b controls neuronal
differentiation by repressing its host transcript, ctdsp2", Genes & development, vol.
26, no. 1, pp. 25-30.
Ding, Q., Dimayuga, E. & Keller, J.N. 2006, "Proteasome regulation of oxidative stress
in aging and age-related diseases of the CNS", Antioxidants & redox signaling, vol.
8, no. 1-2, pp. 163-172.
Dinnis, D.M. & James, D.C. 2005, "Engineering mammalian cell factories for improved
recombinant monoclonal antibody production: lessons from nature?", Biotechnology
and bioengineering, vol. 91, no. 2, pp. 180-189.
Djuranovic, S., Nahvi, A. & Green, R. 2012, "miRNA-mediated gene silencing by
translational repression followed by mRNA deadenylation and decay", Science (New
York, N.Y.), vol. 336, no. 6078, pp. 237-240.
Donald, S.P., Sun, X.Y., Hu, C.A., Yu, J., Mei, J.M., Valle, D. & Phang, J.M. 2001,
"Proline oxidase, encoded by p53-induced gene-6, catalyzes the generation of
proline-dependent reactive oxygen species", Cancer research, vol. 61, no. 5, pp.
1810-1815.
Donaldson, M., Wood, H.A., Kulakosky, P.C. & Shuler, M.L. 1999, "Glycosylation of a
recombinant protein in the Tn5B1-4 insect cell line: influence of ammonia, time of
harvest, temperature, and dissolved oxygen", Biotechnology and bioengineering,
vol. 63, no. 3, pp. 255-262.
375
Dong, C.G., Wu, W.K., Feng, S.Y., Wang, X.J., Shao, J.F. & Qiao, J. 2012, "Co-
inhibition of microRNA-10b and microRNA-21 exerts synergistic inhibition on the
proliferation and invasion of human glioma cells", International journal of
oncology,41(31), pp. 1005-12.
Doolan, P., Clarke, C., Kinsella, P., Breen, L., Meleady, P., Leonard, M., Zhang, L.,
Clynes, M., Aherne, S. & Barron, N. 2013, "Transcriptomic analysis of clonal growth
rate variation during CHO cell line development", Journal of Biotechnology, 10;
166(3), pp. 105-13.
Doolan, P., Meleady, P., Barron, N., Henry, M., Gallagher, R., Gammell, P., Melville,
M., Sinacore, M., McCarthy, K., Leonard, M., Charlebois, T. & Clynes, M. 2010,
"Microarray and proteomics expression profiling identifies several candidates,
including the valosin-containing protein (VCP), involved in regulating high cellular
growth rate in production CHO cell lines", Biotechnology and bioengineering, vol.
106, no. 1, pp. 42-56.
Dowd, J.E., Kwok, K.E. & Piret, J.M. 2000, "Increased t-PA yields using ultrafiltration
of an inhibitory product from CHO fed-batch culture", Biotechnology progress, vol.
16, no. 5, pp. 786-794.
Drissi, R., Dubois, M.L. & Boisvert, F.M. 2013, "Proteomics methods for subcellular
proteome analysis", The FEBS journal, vol. 280, no. 22, pp. 5626-5634.
Druz, A., Chu, C., Majors, B., Santuary, R., Betenbaugh, M. & Shiloach, J. 2011, "A
novel microRNA mmu-miR-466h affects apoptosis regulation in mammalian cells",
Biotechnology and bioengineering, vol. 108, no. 7, pp. 1651-1661.
Druz, A., Son, Y.J., Betenbaugh, M. & Shiloach, J. 2013, "Stable inhibition of mmu-miR-
466h-5p improves apoptosis resistance and protein production in CHO cells",
Metabolic engineering, vol. 16C, pp. 87-94.
Du, G., Yonekubo, J., Zeng, Y., Osisami, M. & Frohman, M.A. 2006, "Design of
expression vectors for RNA interference based on miRNAs and RNA splicing", The
FEBS journal, vol. 273, no. 23, pp. 5421-5427.
Du, R., Sun, W., Xia, L., Zhao, A., Yu, Y., Zhao, L., Wang, H., Huang, C. & Sun, S.
2012, "Hypoxia-induced down-regulation of microRNA-34a promotes EMT by
targeting the Notch signaling pathway in tubular epithelial cells", PloS one, vol. 7,
no. 2, pp. e30771.
Du, W.W., Fang, L., Li, M., Yang, X., Liang, Y., Peng, C., Qian, W., O'Malley, Y.Q.,
Askeland, R.W., Sugg, S.L., Qian, J., Lin, J., Jiang, Z., Yee, A.J., Sefton, M., Deng,
Z., Shan, S.W., Wang, C.H. & Yang, B.B. 2013, "MicroRNA miR-24 enhances
tumor invasion and metastasis by targeting PTPN9 and PTPRF to promote EGF
signaling", Journal of cell science, vol. 126, no. Pt 6, pp. 1440-1453.
Dupuy, A.J., Jenkins, N.A. & Copeland, N.G. 2006, "Sleeping beauty: a novel cancer
gene discovery tool", Human molecular genetics, vol. 15 Spec No 1, pp. R75-9.
376
Dweep, H., Sticht, C., Pandey, P. & Gretz, N. 2011a, "miRWalk--database: prediction of
possible miRNA binding sites by "walking" the genes of three genomes", Journal of
Biomedical Informatics, vol. 44, no. 5, pp. 839-847.
Dweep, H., Sticht, C., Pandey, P. & Gretz, N. 2011b, "miRWalk--database: prediction of
possible miRNA binding sites by "walking" the genes of three genomes", Journal of
Biomedical Informatics, vol. 44, no. 5, pp. 839-847.
Easom, R.A. 1999, "CaM kinase II: a protein kinase with extraordinary talents germane
to insulin exocytosis", Diabetes, vol. 48, no. 4, pp. 675-684.
Ebert, M.S., Neilson, J.R. & Sharp, P.A. 2007, "MicroRNA sponges: competitive
inhibitors of small RNAs in mammalian cells", Nature methods, vol. 4, no. 9, pp.
721-726.
EDELMAN, G.M. & BENACERRAF, B. 1962, "On structural and functional relations
between antibodies and proteins of the gamma-system", Proceedings of the National
Academy of Sciences of the United States of America, vol. 48, pp. 1035-1042.
Ehrenreich, I.M. & Purugganan, M. 2008, "MicroRNAs in plants: Possible contributions
to phenotypic diversity", Plant signaling & behavior, vol. 3, no. 10, pp. 829-830.
Eilers, M. & Eisenman, R.N. 2008, "Myc's broad reach", Genes & development, vol. 22,
no. 20, pp. 2755-2766.
Eizirik, D.L., Miani, M. & Cardozo, A.K. 2013, "Signalling danger: endoplasmic
reticulum stress and the unfolded protein response in pancreatic islet inflammation",
Diabetologia, vol. 56, no. 2, pp. 234-241.
Elkan-Miller, T., Ulitsky, I., Hertzano, R., Rudnicki, A., Dror, A.A., Lenz, D.R., Elkon,
R., Irmler, M., Beckers, J., Shamir, R. & Avraham, K.B. 2011, "Integration of
transcriptomics, proteomics, and microRNA analyses reveals novel microRNA
regulation of targets in the mammalian inner ear", PloS one, vol. 6, no. 4, pp. e18195.
Eng, K.E., Panas, M.D., Karlsson Hedestam, G.B. & McInerney, G.M. 2010, "A novel
quantitative flow cytometry-based assay for autophagy", Autophagy, vol. 6, no. 5,
pp. 634-641.
Esau, C., Davis, S., Murray, S.F., Yu, X.X., Pandey, S.K., Pear, M., Watts, L., Booten,
S.L., Graham, M., McKay, R., Subramaniam, A., Propp, S., Lollo, B.A., Freier, S.,
Bennett, C.F., Bhanot, S. & Monia, B.P. 2006a, "miR-122 regulation of lipid
metabolism revealed by in vivo antisense targeting", Cell metabolism, vol. 3, no. 2,
pp. 87-98.
Esau, C., Davis, S., Murray, S.F., Yu, X.X., Pandey, S.K., Pear, M., Watts, L., Booten,
S.L., Graham, M., McKay, R., Subramaniam, A., Propp, S., Lollo, B.A., Freier, S.,
Bennett, C.F., Bhanot, S. & Monia, B.P. 2006b, "miR-122 regulation of lipid
metabolism revealed by in vivo antisense targeting", Cell metabolism, vol. 3, no. 2,
pp. 87-98.
377
Ferl, R.J., Manak, M.S. & Reyes, M.F. 2002, "The 14-3-3s", Genome biology, vol. 3, no.
7, pp. REVIEWS3010.
Fernando, M.R., Nanri, H., Yoshitake, S., Nagata-Kuno, K. & Minakami, S. 1992,
"Thioredoxin regenerates proteins inactivated by oxidative stress in endothelial
cells", European journal of biochemistry / FEBS, vol. 209, no. 3, pp. 917-922.
Ferrara, C., Brunker, P., Suter, T., Moser, S., Puntener, U. & Umana, P. 2006,
"Modulation of therapeutic antibody effector functions by glycosylation
engineering: influence of Golgi enzyme localization domain and co-expression of
heterologous beta1, 4-N-acetylglucosaminyltransferase III and Golgi alpha-
mannosidase II", Biotechnology and bioengineering, vol. 93, no. 5, pp. 851-861.
Figueroa, B.,Jr, Chen, S., Oyler, G.A., Hardwick, J.M. & Betenbaugh, M.J. 2004, "Aven
and Bcl-xL enhance protection against apoptosis for mammalian cells exposed to
various culture conditions", Biotechnology and bioengineering, vol. 85, no. 6, pp.
589-600.
Filipowicz, W., Bhattacharyya, S.N. & Sonenberg, N. 2008, "Mechanisms of post-
transcriptional regulation by microRNAs: are the answers in sight?", Nature
reviews.Genetics, vol. 9, no. 2, pp. 102-114.
Fiore, M., Zanier, R. & Degrassi, F. 2002, "Reversible G(1) arrest by dimethyl sulfoxide
as a new method to synchronize Chinese hamster cells", Mutagenesis, vol. 17, no. 5,
pp. 419-424.
Fischer, S., Wagner, A., Kos, A., Aschrafi, A., Handrick, R., Hannemann, J. & Otte, K.
2013, "Breaking limitations of complex culture media: functional non-viral miRNA
delivery into pharmaceutical production cell lines", Journal of Biotechnology, vol.
168, no. 4, pp. 589-600.
Flowers, E., Froelicher, E.S. & Aouizerat, B.E. 2012, "MicroRNA regulation of lipid
metabolism", Metabolism: clinical and experimental, 62(1), pp. 12-20.
Ford, L.P. 2006, "Using synthetic miRNA mimics for diverting cell fate: a possibility of
miRNA-based therapeutics?", Leukemia research, vol. 30, no. 5, pp. 511-513.
Fornuskova, D., Stiburek, L., Wenchich, L., Vinsova, K., Hansikova, H. & Zeman, J.
2010, "Novel insights into the assembly and function of human nuclear-encoded
cytochrome c oxidase subunits 4, 5a, 6a, 7a and 7b", The Biochemical journal, vol.
428, no. 3, pp. 363-374.
Forstemann, K., Tomari, Y., Du, T., Vagin, V.V., Denli, A.M., Bratu, D.P., Klattenhoff,
C., Theurkauf, W.E. & Zamore, P.D. 2005, "Normal microRNA maturation and
germ-line stem cell maintenance requires Loquacious, a double-stranded RNA-
binding domain protein", PLoS biology, vol. 3, no. 7, pp. e236.
Fox, S.R., Patel, U.A., Yap, M.G. & Wang, D.I. 2004, "Maximizing interferon-gamma
production by Chinese hamster ovary cells through temperature shift optimization:
378
experimental and modeling", Biotechnology and bioengineering, vol. 85, no. 2, pp.
177-184.
Franco-Zorrilla, J.M., Valli, A., Todesco, M., Mateos, I., Puga, M.I., Rubio-Somoza, I.,
Leyva, A., Weigel, D., Garcia, J.A. & Paz-Ares, J. 2007, "Target mimicry provides
a new mechanism for regulation of microRNA activity", Nature genetics, vol. 39,
no. 8, pp. 1033-1037.
Freedman, R.B., Hirst, T.R. & Tuite, M.F. 1994, "Protein disulphide isomerase: building
bridges in protein folding", Trends in biochemical sciences, vol. 19, no. 8, pp. 331-
336.
Friday, E., Oliver, R., Turturro, F. & Welbourne, T. 2012, "Role of glutamate
dehydrogenase in cancer growth and homeostasis", Dehydrogenases.Rijeka,
Croatia: InTech, pp. 181-190.
Friedman, R.C., Farh, K.K., Burge, C.B. & Bartel, D.P. 2009, "Most mammalian mRNAs
are conserved targets of microRNAs", Genome research, vol. 19, no. 1, pp. 92-105.
Frost, R.J. & Olson, E.N. 2011, "Control of glucose homeostasis and insulin sensitivity
by the Let-7 family of microRNAs", Proceedings of the National Academy of
Sciences of the United States of America, vol. 108, no. 52, pp. 21075-21080.
Fujita, Y., Kojima, K., Hamada, N., Ohhashi, R., Akao, Y., Nozawa, Y., Deguchi, T. &
Ito, M. 2008, "Effects of miR-34a on cell growth and chemoresistance in prostate
cancer PC3 cells", Biochemical and biophysical research communications, vol. 377,
no. 1, pp. 114-119.
Furukawa, N., Sakurai, F., Katayama, K., Seki, N., Kawabata, K. & Mizuguchi, H. 2011,
"Optimization of a microRNA expression vector for function analysis of
microRNA", Journal of controlled release : official journal of the Controlled
Release Society, vol. 150, no. 1, pp. 94-101.
Fussenegger, M., Mazur, X. & Bailey, J.E. 1997, "A novel cytostatic process enhances
the productivity of Chinese hamster ovary cells", Biotechnology and bioengineering,
vol. 55, no. 6, pp. 927-939.
Gammell, P., Barron, N., Kumar, N. & Clynes, M. 2007, "Initial identification of low
temperature and culture stage induction of miRNA expression in suspension CHO-
K1 cells", Journal of Biotechnology, vol. 130, no. 3, pp. 213-218.
Gao, P., Sun, L., He, X., Cao, Y. & Zhang, H. 2012, "MicroRNAs and the Warburg
Effect: new players in an old arena", Current gene therapy, vol. 12, no. 4, pp. 285-
291.
Gao, P., Tchernyshyov, I., Chang, T.C., Lee, Y.S., Kita, K., Ochi, T., Zeller, K.I., De
Marzo, A.M., Van Eyk, J.E., Mendell, J.T. & Dang, C.V. 2009, "c-Myc suppression
of miR-23a/b enhances mitochondrial glutaminase expression and glutamine
metabolism", Nature, vol. 458, no. 7239, pp. 762-765.
379
Garcia, D. 2008, "A miRacle in plant development: role of microRNAs in cell
differentiation and patterning", Seminars in cell & developmental biology, vol. 19,
no. 6, pp. 586-595.
Garcia, D.M., Baek, D., Shin, C., Bell, G.W., Grimson, A. & Bartel, D.P. 2011, "Weak
seed-pairing stability and high target-site abundance decrease the proficiency of lsy-
6 and other microRNAs", Nature structural & molecular biology, vol. 18, no. 10, pp.
1139-1146.
Gardino, A.K., Smerdon, S.J. & Yaffe, M.B. 2006, "Structural determinants of 14-3-3
binding specificities and regulation of subcellular localization of 14-3-3-ligand
complexes: a comparison of the X-ray crystal structures of all human 14-3-3
isoforms", Seminars in cancer biology, vol. 16, no. 3, pp. 173-182.
Gaziel-Sovran, A. & Hernando, E. 2012, "miRNA-mediated GALNT modulation of
invasion and immune suppression: A sweet deal for metastatic cells",
Oncoimmunology, vol. 1, no. 5, pp. 746-748.
Gellerich, F.N., Gizatullina, Z., Trumbeckaite, S., Nguyen, H.P., Pallas, T.,
Arandarcikaite, O., Vielhaber, S., Seppet, E. & Striggow, F. 2010, "The regulation
of OXPHOS by extramitochondrial calcium", Biochimica et biophysica acta, vol.
1797, no. 6-7, pp. 1018-1027.
Georgantas, R.W.,3rd, Streicher, K., Luo, X., Greenlees, L., Zhu, W., Liu, Z., Brohawn,
P., Morehouse, C., Higgs, B.W., Richman, L., Jallal, B., Yao, Y. & Ranade, K. 2014,
"MicroRNA-206 induces G1 arrest in melanoma by inhibition of CDK4 and Cyclin
D", Pigment cell & melanoma research, vol. 27, no. 2, pp. 275-286.
Gething, M.J. & Sambrook, J. 1992, "Protein folding in the cell", Nature, vol. 355, no.
6355, pp. 33-45.
Ghorbaniaghdam, A., Henry, O. & Jolicoeur, M. 2013, "A kinetic-metabolic model based
on cell energetic state: study of CHO cell behavior under Na-butyrate stimulation",
Bioprocess and biosystems engineering, vol. 36, no. 4, pp. 469-487.
Giglio, S., Cirombella, R., Amodeo, R., Portaro, L., Lavra, L. & Vecchione, A. 2013,
"MicroRNA miR-24 promotes cell proliferation by targeting the CDKs inhibitors
p27 and p16", Journal of cellular physiology, 228(10), pp. 2015-23.
Gilbert, M.R., Liu, Y., Neltner, J., Pu, H., Morris, A., Sunkara, M., Pittman, T.,
Kyprianou, N. & Horbinski, C. 2014, "Autophagy and oxidative stress in gliomas
with IDH1 mutations", Acta Neuropathologica, vol. 127, no. 2, pp. 221-233.
Gill, S.S., Anjum, N.A., Hasanuzzaman, M., Gill, R., Trivedi, D.K., Ahmad, I., Pereira,
E. & Tuteja, N. 2013, "Glutathione and glutathione reductase: a boon in disguise for
plant abiotic stress defense operations", Plant Physiology and Biochemistry : PPB /
Societe francaise de physiologie vegetale, vol. 70, pp. 204-212.
Godlewski, J., Nowicki, M.O., Bronisz, A., Nuovo, G., Palatini, J., De Lay, M., Van
Brocklyn, J., Ostrowski, M.C., Chiocca, E.A. & Lawler, S.E. 2010, "MicroRNA-451
380
regulates LKB1/AMPK signaling and allows adaptation to metabolic stress in glioma
cells", Molecular cell, vol. 37, no. 5, pp. 620-632.
Griffiths-Jones, S. 2004, "The microRNA Registry", Nucleic acids research, vol. 32, no.
Database issue, pp. D109-11.
Griffiths-Jones, S., Hui, J.H., Marco, A. & Ronshaugen, M. 2011, "MicroRNA evolution
by arm switching", EMBO reports, vol. 12, no. 2, pp. 172-177.
Griffiths-Jones, S., Saini, H.K., van Dongen, S. & Enright, A.J. 2008, "miRBase: tools
for microRNA genomics", Nucleic acids research, vol. 36, no. Database issue, pp.
D154-8.
Grimm, D., Streetz, K.L., Jopling, C.L., Storm, T.A., Pandey, K., Davis, C.R., Marion,
P., Salazar, F. & Kay, M.A. 2006, "Fatality in mice due to oversaturation of cellular
microRNA/short hairpin RNA pathways", Nature, vol. 441, no. 7092, pp. 537-541.
Grimson, A., Farh, K.K., Johnston, W.K., Garrett-Engele, P., Lim, L.P. & Bartel, D.P.
2007, "MicroRNA targeting specificity in mammals: determinants beyond seed
pairing", Molecular cell, vol. 27, no. 1, pp. 91-105.
Grozinger, C.M. & Schreiber, S.L. 2000, "Regulation of histone deacetylase 4 and 5 and
transcriptional activity by 14-3-3-dependent cellular localization", Proceedings of
the National Academy of Sciences of the United States of America, vol. 97, no. 14,
pp. 7835-7840.
Grune, T., Catalgol, B., Licht, A., Ermak, G., Pickering, A.M., Ngo, J.K. & Davies, K.J.
2011, "HSP70 mediates dissociation and reassociation of the 26S proteasome during
adaptation to oxidative stress", Free radical biology & medicine, vol. 51, no. 7, pp.
1355-1364.
Gu, S., Jin, L., Zhang, F., Huang, Y., Grimm, D., Rossi, J.J. & Kay, M.A. 2011,
"Thermodynamic stability of small hairpin RNAs highly influences the loading
process of different mammalian Argonautes", Proceedings of the National Academy
of Sciences of the United States of America, vol. 108, no. 22, pp. 9208-9213.
Guan, Y., Chen, Q., Yang, X., Haines, P., Pei, M., Terek, R., Wei, X., Zhao, T. & Wei,
L. 2012, "Subcellular relocation of histone deacetylase 4 regulates growth plate
chondrocyte differentiation through Ca2+/calmodulin-dependent kinase IV",
American journal of physiology.Cell physiology, vol. 303, no. 1, pp. C33-40.
Guo, L. & Lu, Z. 2010, "The fate of miRNA* strand through evolutionary analysis:
implication for degradation as merely carrier strand or potential regulatory
molecule?", PloS one, vol. 5, no. 6, pp. e11387.
Haberland, M., Montgomery, R.L. & Olson, E.N. 2009, "The many roles of histone
deacetylases in development and physiology: implications for disease and therapy",
Nature reviews.Genetics, vol. 10, no. 1, pp. 32-42.
381
Hackl, M., Borth, N. & Grillari, J. 2012, "miRNAs--pathway engineering of CHO cell
factories that avoids translational burdening", Trends in biotechnology, vol. 30, no.
8, pp. 405-406.
Hackl, M., Jadhav, V., Jakobi, T., Rupp, O., Brinkrolf, K., Goesmann, A., Puhler, A.,
Noll, T., Borth, N. & Grillari, J. 2012, "Computational identification of microRNA
gene loci and precursor microRNA sequences in CHO cell lines", Journal of
Biotechnology, vol. 158, no. 3, pp. 151-155.
Hackl, M., Jadhav, V., Klanert, G., Karbiener, M., Scheideler, M., Grillari, J. & Borth, N.
2014a, "Analysis of microRNA transcription and post-transcriptional processing by
Dicer in the context of CHO cell proliferation", Journal of Biotechnology, 2014 Jan
28. pii: S0168-1656(14)00037-6. doi: 10.1016/j.jbiotec.2013.12.018.
Hackl, M., Jakobi, T., Blom, J., Doppmeier, D., Brinkrolf, K., Szczepanowski, R.,
Bernhart, S.H., Honer Zu Siederdissen, C., Bort, J.A., Wieser, M., Kunert, R., Jeffs,
S., Hofacker, I.L., Goesmann, A., Puhler, A., Borth, N. & Grillari, J. 2011, "Next-
generation sequencing of the Chinese hamster ovary microRNA transcriptome:
Identification, annotation and profiling of microRNAs as targets for cellular
engineering", Journal of Biotechnology, vol. 153, no. 1-2, pp. 62-75.
Hammell, M., Long, D., Zhang, L., Lee, A., Carmack, C.S., Han, M., Ding, Y. & Ambros,
V. 2008, "mirWIP: microRNA target prediction based on microRNA-containing
ribonucleoprotein-enriched transcripts", Nature methods, vol. 5, no. 9, pp. 813-819.
Hammond, S., Kaplarevic, M., Borth, N., Betenbaugh, M.J. & Lee, K.H. 2012, "Chinese
hamster genome database: an online resource for the CHO community at
www.CHOgenome.org", Biotechnology and bioengineering, vol. 109, no. 6, pp.
1353-1356.
Han, B.W., Hung, J.H., Weng, Z., Zamore, P.D. & Ameres, S.L. 2011, "The 3'-to-5'
exoribonuclease Nibbler shapes the 3' ends of microRNAs bound to Drosophila
Argonaute1", Current biology : CB, vol. 21, no. 22, pp. 1878-1887.
Han, J., Lee, Y., Yeom, K.H., Kim, Y.K., Jin, H. & Kim, V.N. 2004, "The Drosha-
DGCR8 complex in primary microRNA processing", Genes & development, vol. 18,
no. 24, pp. 3016-3027.
Han, J., Lee, Y., Yeom, K.H., Nam, J.W., Heo, I., Rhee, J.K., Sohn, S.Y., Cho, Y., Zhang,
B.T. & Kim, V.N. 2006, "Molecular basis for the recognition of primary microRNAs
by the Drosha-DGCR8 complex", Cell, vol. 125, no. 5, pp. 887-901.
Han, Y.K., Ha, T.K., Lee, S.J., Lee, J.S. & Lee, G.M. 2011, "Autophagy and apoptosis of
recombinant Chinese hamster ovary cells during fed-batch culture: effect of nutrient
supplementation", Biotechnology and bioengineering, vol. 108, no. 9, pp. 2182-
2192.
Hansen, T.B., Wiklund, E.D., Bramsen, J.B., Villadsen, S.B., Statham, A.L., Clark, S.J.
& Kjems, J. 2011, "miRNA-dependent gene silencing involving Ago2-mediated
382
cleavage of a circular antisense RNA", The EMBO journal, vol. 30, no. 21, pp. 4414-
4422.
Haraguchi, T., Ozaki, Y. & Iba, H. 2009, "Vectors expressing efficient RNA decoys
achieve the long-term suppression of specific microRNA activity in mammalian
cells", Nucleic acids research, vol. 37, no. 6, pp. e43.
Harper, M.E., Dent, R., Monemdjou, S., Bezaire, V., Van Wyck, L., Wells, G., Kavaslar,
G.N., Gauthier, A., Tesson, F. & McPherson, R. 2002, "Decreased mitochondrial
proton leak and reduced expression of uncoupling protein 3 in skeletal muscle of
obese diet-resistant women", Diabetes, vol. 51, no. 8, pp. 2459-2466.
Hashimi, S.T., Fulcher, J.A., Chang, M.H., Gov, L., Wang, S. & Lee, B. 2009,
"MicroRNA profiling identifies miR-34a and miR-21 and their target genes JAG1
and WNT1 in the coordinate regulation of dendritic cell differentiation", Blood, vol.
114, no. 2, pp. 404-414.
Hassan, T., Smith, S.G., Gaughan, K., Oglesby, I.K., O'Neill, S., McElvaney, N.G. &
Greene, C.M. 2013, "Isolation and identification of cell-specific microRNAs
targeting a messenger RNA using a biotinylated anti-sense oligonucleotide capture
affinity technique", Nucleic acids research, vol. 41, no. 6, pp. e71.
He, L., He, X., Lim, L.P., de Stanchina, E., Xuan, Z., Liang, Y., Xue, W., Zender, L.,
Magnus, J., Ridzon, D., Jackson, A.L., Linsley, P.S., Chen, C., Lowe, S.W., Cleary,
M.A. & Hannon, G.J. 2007a, "A microRNA component of the p53 tumour
suppressor network", Nature, vol. 447, no. 7148, pp. 1130-1134.
He, L., He, X., Lim, L.P., de Stanchina, E., Xuan, Z., Liang, Y., Xue, W., Zender, L.,
Magnus, J., Ridzon, D., Jackson, A.L., Linsley, P.S., Chen, C., Lowe, S.W., Cleary,
M.A. & Hannon, G.J. 2007b, "A microRNA component of the p53 tumour
suppressor network", Nature, vol. 447, no. 7148, pp. 1130-1134.
He, L., He, X., Lim, L.P., de Stanchina, E., Xuan, Z., Liang, Y., Xue, W., Zender, L.,
Magnus, J., Ridzon, D., Jackson, A.L., Linsley, P.S., Chen, C., Lowe, S.W., Cleary,
M.A. & Hannon, G.J. 2007c, "A microRNA component of the p53 tumour
suppressor network", Nature, vol. 447, no. 7148, pp. 1130-1134.
Healy, S., Khan, D.H. & Davie, J.R. 2011, "Gene expression regulation through 14-3-3
interactions with histones and HDACs", Discovery medicine, vol. 11, no. 59, pp.
349-358.
Hendrick, V., Winnepenninckx, P., Abdelkafi, C., Vandeputte, O., Cherlet, M., Marique,
T., Renemann, G., Loa, A., Kretzmer, G. & Werenne, J. 2001a, "Increased
productivity of recombinant tissular plasminogen activator (t-PA) by butyrate and
shift of temperature: a cell cycle phases analysis", Cytotechnology, vol. 36, no. 1-3,
pp. 71-83.
Hendrick, V., Winnepenninckx, P., Abdelkafi, C., Vandeputte, O., Cherlet, M., Marique,
T., Renemann, G., Loa, A., Kretzmer, G. & Werenne, J. 2001b, "Increased
productivity of recombinant tissular plasminogen activator (t-PA) by butyrate and
383
shift of temperature: a cell cycle phases analysis", Cytotechnology, vol. 36, no. 1-3,
pp. 71-83.
Hennessy, E., Clynes, M., Jeppesen, P.B. & O'Driscoll, L. 2010, "Identification of
microRNAs with a role in glucose stimulated insulin secretion by expression
profiling of MIN6 cells", Biochemical and biophysical research communications,
vol. 396, no. 2, pp. 457-462.
Hermeking, H. 2009, "MiR-34a and p53", Cell cycle (Georgetown, Tex.), vol. 8, no. 9,
pp. 1308.
Hermeking, H. 2007, "p53 enters the microRNA world", Cancer cell, vol. 12, no. 5, pp.
414-418.
Hernandez Bort, J.A., Hackl, M., Hoflmayer, H., Jadhav, V., Harreither, E., Kumar, N.,
Ernst, W., Grillari, J. & Borth, N. 2012, "Dynamic mRNA and miRNA profiling of
CHO-K1 suspension cell cultures", Biotechnology journal, vol. 7, no. 4, pp. 500-
515.
Hirano, N., Muroi, T., Takahashi, H. & Haruki, M. 2011, "Site-specific recombinases as
tools for heterologous gene integration", Applied Microbiology and Biotechnology,
vol. 92, no. 2, pp. 227-239.
Hirsch, A., Hahn, D., Kempna, P., Hofer, G., Mullis, P.E., Nuoffer, J.M. & Fluck, C.E.
2012, "Role of AMP-activated protein kinase on steroid hormone biosynthesis in
adrenal NCI-H295R cells", PloS one, vol. 7, no. 1, pp. e30956.
Ho, S.C., Tong, Y.W. & Yang, Y. 2013, "Generation of monoclonal antibody-producing
mammalian cell lines", Pharmaceutical Bioprocessing, vol. 1, no. 1, pp. 71-87.
Hong, J.K., Cho, S.M. & Yoon, S.K. 2010, "Substitution of glutamine by glutamate
enhances production and galactosylation of recombinant IgG in Chinese hamster
ovary cells", Applied Microbiology and Biotechnology, vol. 88, no. 4, pp. 869-876.
Horbinski, C. 2013, "What do we know about IDH1/2 mutations so far, and how do we
use it?", Acta Neuropathologica, vol. 125, no. 5, pp. 621-636.
Hsieh, J.K., Fredersdorf, S., Kouzarides, T., Martin, K. & Lu, X. 1997, "E2F1-induced
apoptosis requires DNA binding but not transactivation and is inhibited by the
retinoblastoma protein through direct interaction", Genes & development, vol. 11,
no. 14, pp. 1840-1852.
Hu, H.Y., Yan, Z., Xu, Y., Hu, H., Menzel, C., Zhou, Y.H., Chen, W. & Khaitovich, P.
2009, "Sequence features associated with microRNA strand selection in humans and
flies", BMC genomics, vol. 10, pp. 413.
Huang, Y.M., Hu, W., Rustandi, E., Chang, K., Yusuf-Makagiansar, H. & Ryll, T. 2010,
"Maximizing productivity of CHO cell-based fed-batch culture using chemically
defined media conditions and typical manufacturing equipment", Biotechnology
progress, vol. 26, no. 5, pp. 1400-1410.
384
Hwang, S.O., Chung, J.Y. & Lee, G.M. 2003, "Effect of doxycycline-regulated ERp57
expression on specific thrombopoietin productivity of recombinant CHO cells",
Biotechnology progress, vol. 19, no. 1, pp. 179-184.
Hwang, S.O. & Lee, G.M. 2009, "Effect of Akt overexpression on programmed cell death
in antibody-producing Chinese hamster ovary cells", Journal of Biotechnology, vol.
139, no. 1, pp. 89-94.
Hwang, S.O. & Lee, G.M. 2008, "Nutrient deprivation induces autophagy as well as
apoptosis in Chinese hamster ovary cell culture", Biotechnology and bioengineering,
vol. 99, no. 3, pp. 678-685.
Ichimura, A., Ruike, Y., Terasawa, K., Shimizu, K. & Tsujimoto, G. 2010, "MicroRNA-
34a inhibits cell proliferation by repressing mitogen-activated protein kinase kinase
1 during megakaryocytic differentiation of K562 cells", Molecular pharmacology,
vol. 77, no. 6, pp. 1016-1024.
Ikeda, Y., Tanji, E., Makino, N., Kawata, S. & Furukawa, T. 2012, "MicroRNAs
associated with mitogen-activated protein kinase in human pancreatic cancer",
Molecular cancer research : MCR, vol. 10, no. 2, pp. 259-269.
Imai-Nishiya, H., Mori, K., Inoue, M., Wakitani, M., Iida, S., Shitara, K. & Satoh, M.
2007, "Double knockdown of alpha1,6-fucosyltransferase (FUT8) and GDP-
mannose 4,6-dehydratase (GMD) in antibody-producing cells: a new strategy for
generating fully non-fucosylated therapeutic antibodies with enhanced ADCC",
BMC biotechnology, vol. 7, pp. 84.
Iqbal, M.A., Gupta, V., Gopinath, P., Mazurek, S. & Bamezai, R.N. 2014, "Pyruvate
kinase M2 and cancer: an updated assessment", FEBS letters, 19;588(16), pp. 2685-
92.
Ishteiwy, R.A., Ward, T.M., Dykxhoorn, D.M. & Burnstein, K.L. 2012, "The microRNA
-23b/-27b cluster suppresses the metastatic phenotype of castration-resistant prostate
cancer cells", PloS one, vol. 7, no. 12, pp. e52106.
Itani, S., Kunisada, T., Morimoto, Y., Yoshida, A., Sasaki, T., Ito, S., Ouchida, M.,
Sugihara, S., Shimizu, K. & Ozaki, T. 2012, "MicroRNA-21 correlates with
tumorigenesis in malignant peripheral nerve sheath tumor (MPNST) via
programmed cell death protein 4 (PDCD4)", Journal of cancer research and clinical
oncology, vol. 138, no. 9, pp. 1501-1509.
Ivanovska, I., Ball, A.S., Diaz, R.L., Magnus, J.F., Kibukawa, M., Schelter, J.M.,
Kobayashi, S.V., Lim, L., Burchard, J., Jackson, A.L., Linsley, P.S. & Cleary, M.A.
2008, "MicroRNAs in the miR-106b family regulate p21/CDKN1A and promote cell
cycle progression", Molecular and cellular biology, vol. 28, no. 7, pp. 2167-2174.
Jacobson, J. & Duchen, M.R. 2004, "Interplay between mitochondria and cellular calcium
signalling", Molecular and cellular biochemistry, vol. 256-257, no. 1-2, pp. 209-218.
385
Jadhav, V., Hackl, M., Bort, J.A., Wieser, M., Harreither, E., Kunert, R., Borth, N. &
Grillari, J. 2012, "A screening method to assess biological effects of microRNA
overexpression in Chinese hamster ovary cells", Biotechnology and bioengineering,
vol. 109, no. 6, pp. 1376-1385.
Jadhav, V., Hackl, M., Druz, A., Shridhar, S., Chung, C.Y., Heffner, K.M., Kreil, D.P.,
Betenbaugh, M., Shiloach, J., Barron, N., Grillari, J. & Borth, N. 2013, "CHO
microRNA engineering is growing up: Recent successes and future challenges",
Biotechnology Advances, 2013 Dec;31(8):1501-13. doi:
10.1016/j.biotechadv.2013.07.00.
Jadhav, V., Hackl, M., Klanert, G., Hernandez Bort, J.A., Kunert, R., Grillari, J. & Borth,
N. 2014, "Stable overexpression of miR-17 enhances recombinant protein
production of CHO cells", Journal of Biotechnology, vol. 175C, pp. 38-44.
Jahn, R. & Scheller, R.H. 2006, "SNAREs--engines for membrane fusion", Nature
reviews.Molecular cell biology, vol. 7, no. 9, pp. 631-643.
Jaluria, P., Betenbaugh, M., Konstantopoulos, K. & Shiloach, J. 2007, "Enhancement of
cell proliferation in various mammalian cell lines by gene insertion of a cyclin-
dependent kinase homolog", BMC biotechnology, vol. 7, pp. 71.
Janas, M.M., Wang, B., Harris, A.S., Aguiar, M., Shaffer, J.M., Subrahmanyam, Y.V.,
Behlke, M.A., Wucherpfennig, K.W., Gygi, S.P., Gagnon, E. & Novina, C.D. 2012,
"Alternative RISC assembly: binding and repression of microRNA-mRNA duplexes
by human Ago proteins", RNA (New York, N.Y.), vol. 18, no. 11, pp. 2041-2055.
Jastroch, M., Divakaruni, A.S., Mookerjee, S., Treberg, J.R. & Brand, M.D. 2010,
"Mitochondrial proton and electron leaks", Essays in biochemistry, vol. 47, pp. 53-
67.
Jayapal, K.P., Wlaschin, K.F., Hu, W. & Yap, M.G.S. 2007, "Recombinant protein
therapeutics from CHO cells-20 years and counting", Chemical Engineering
Progress, vol. 103, no. 10, pp. 40.
Jeck, W.R., Sorrentino, J.A., Wang, K., Slevin, M.K., Burd, C.E., Liu, J., Marzluff, W.F.
& Sharpless, N.E. 2013, "Circular RNAs are abundant, conserved, and associated
with ALU repeats", RNA (New York, N.Y.), vol. 19, no. 2, pp. 141-157.
Jefferis, R. 2009a, "Glycosylation as a strategy to improve antibody-based therapeutics",
Nature reviews.Drug discovery, vol. 8, no. 3, pp. 226-234.
Jefferis, R. 2009b, "Recombinant antibody therapeutics: the impact of glycosylation on
mechanisms of action", Trends in pharmacological sciences, vol. 30, no. 7, pp. 356-
362.
Jefferis, R. 2007, "Antibody therapeutics: isotype and glycoform selection", Expert
opinion on biological therapy, vol. 7, no. 9, pp. 1401-1413.
386
Jeon, M.K., Yu, D.Y. & Lee, G.M. 2011, "Combinatorial engineering of ldh-a and bcl-2
for reducing lactate production and improving cell growth in dihydrofolate
reductase-deficient Chinese hamster ovary cells", Applied Microbiology and
Biotechnology, vol. 92, no. 4, pp. 779-790.
Jeong, D., Kim, T.S., Lee, J.W., Kim, K.T., Kim, H.J., Kim, I.H. & Kim, I.Y. 2001,
"Blocking of acidosis-mediated apoptosis by a reduction of lactate dehydrogenase
activity through antisense mRNA expression", Biochemical and biophysical
research communications, vol. 289, no. 5, pp. 1141-1149.
Jiang, D., Zhao, L. & Clapham, D.E. 2009, "Genome-wide RNAi screen identifies Letm1
as a mitochondrial Ca2+/H+ antiporter", Science (New York, N.Y.), vol. 326, no.
5949, pp. 144-147.
Jiang, D., Zhao, L., Clish, C.B. & Clapham, D.E. 2013a, "Letm1, the mitochondrial
Ca2+/H+ antiporter, is essential for normal glucose metabolism and alters brain
function in Wolf-Hirschhorn syndrome", Proceedings of the National Academy of
Sciences of the United States of America, vol. 110, no. 24, pp. E2249-54.
Jiang, J., Yang, J., Wang, Z., Wu, G. & Liu, F. 2013b, "TFAM is directly regulated by
miR-23b in glioma", Oncology reports, vol. 30, no. 5, pp. 2105-2110.
Jiang, Q., Wang, Y., Hao, Y., Juan, L., Teng, M., Zhang, X., Li, M., Wang, G. & Liu, Y.
2009, "miR2Disease: a manually curated database for microRNA deregulation in
human disease", Nucleic acids research, vol. 37, no. Database issue, pp. D98-104.
Jiang, X.R., Song, A., Bergelson, S., Arroll, T., Parekh, B., May, K., Chung, S., Strouse,
R., Mire-Sluis, A. & Schenerman, M. 2011, "Advances in the assessment and control
of the effector functions of therapeutic antibodies", Nature reviews.Drug discovery,
vol. 10, no. 2, pp. 101-111.
Jiang, Z. & Sharfstein, S.T. 2008, "Sodium butyrate stimulates monoclonal antibody
over-expression in CHO cells by improving gene accessibility", Biotechnology and
bioengineering, vol. 100, no. 1, pp. 189-194.
Jo, D.H., Kim, J.H., Park, W.Y., Kim, K.W., Yu, Y.S. & Kim, J.H. 2011, "Differential
profiles of microRNAs in retinoblastoma cell lines of different proliferation and
adherence patterns", Journal of pediatric hematology/oncology, vol. 33, no. 7, pp.
529-533.
Jo, S.H., Son, M.K., Koh, H.J., Lee, S.M., Song, I.H., Kim, Y.O., Lee, Y.S., Jeong, K.S.,
Kim, W.B., Park, J.W., Song, B.J. & Huh, T.L. 2001, "Control of mitochondrial
redox balance and cellular defense against oxidative damage by mitochondrial
NADP+-dependent isocitrate dehydrogenase", The Journal of biological chemistry,
vol. 276, no. 19, pp. 16168-16176.
John, B., Enright, A.J., Aravin, A., Tuschl, T., Sander, C. & Marks, D.S. 2004, "Human
MicroRNA targets", PLoS biology, vol. 2, no. 11, pp. e363.
387
Johnson, K.C., Jacob, N.M., Nissom, P.M., Hackl, M., Lee, L.H., Yap, M. & Hu, W.S.
2011, "Conserved microRNAs in Chinese hamster ovary cell lines", Biotechnology
and bioengineering, vol. 108, no. 2, pp. 475-480.
Joseph, J.W., Jensen, M.V., Ilkayeva, O., Palmieri, F., Alarcon, C., Rhodes, C.J. &
Newgard, C.B. 2006, "The mitochondrial citrate/isocitrate carrier plays a regulatory
role in glucose-stimulated insulin secretion", The Journal of biological chemistry,
vol. 281, no. 47, pp. 35624-35632.
Joung, J.K. & Sander, J.D. 2013, "TALENs: a widely applicable technology for targeted
genome editing", Nature reviews.Molecular cell biology, vol. 14, no. 1, pp. 49-55.
Junttila, T.T., Parsons, K., Olsson, C., Lu, Y., Xin, Y., Theriault, J., Crocker, L., Pabonan,
O., Baginski, T., Meng, G., Totpal, K., Kelley, R.F. & Sliwkowski, M.X. 2010,
"Superior in vivo efficacy of afucosylated trastuzumab in the treatment of HER2-
amplified breast cancer", Cancer research, vol. 70, no. 11, pp. 4481-4489.
Juvvuna, P.K., Khandelia, P., Lee, L.M. & Makeyev, E.V. 2012, "Argonaute identity
defines the length of mature mammalian microRNAs", Nucleic acids research,
40(14), pp. 6808-20.
Kabeya, Y., Mizushima, N., Ueno, T., Yamamoto, A., Kirisako, T., Noda, T., Kominami,
E., Ohsumi, Y. & Yoshimori, T. 2000, "LC3, a mammalian homologue of yeast
Apg8p, is localized in autophagosome membranes after processing", The EMBO
journal, vol. 19, no. 21, pp. 5720-5728.
Kaller, M., Liffers, S.T., Oeljeklaus, S., Kuhlmann, K., Roh, S., Hoffmann, R.,
Warscheid, B. & Hermeking, H. 2011a, "Genome-wide characterization of miR-34a
induced changes in protein and mRNA expression by a combined pulsed SILAC and
microarray analysis", Molecular & cellular proteomics : MCP, vol. 10, no. 8, pp.
M111.010462.
Kaller, M., Liffers, S.T., Oeljeklaus, S., Kuhlmann, K., Roh, S., Hoffmann, R.,
Warscheid, B. & Hermeking, H. 2011b, "Genome-wide characterization of miR-34a
induced changes in protein and mRNA expression by a combined pulsed SILAC and
microarray analysis", Molecular & cellular proteomics : MCP, vol. 10, no. 8, pp.
M111.010462.
Kameyama, Y., Kawabe, Y., Ito, A. & Kamihira, M. 2010, "An accumulative site-specific
gene integration system using Cre recombinase-mediated cassette exchange",
Biotechnology and bioengineering, vol. 105, no. 6, pp. 1106-1114.
Kanda, Y., Imai-Nishiya, H., Kuni-Kamochi, R., Mori, K., Inoue, M., Kitajima-Miyama,
K., Okazaki, A., Iida, S., Shitara, K. & Satoh, M. 2007, "Establishment of a GDP-
mannose 4,6-dehydratase (GMD) knockout host cell line: a new strategy for
generating completely non-fucosylated recombinant therapeutics", Journal of
Biotechnology, vol. 130, no. 3, pp. 300-310.
Kanda, Y., Yamane-Ohnuki, N., Sakai, N., Yamano, K., Nakano, R., Inoue, M., Misaka,
H., Iida, S., Wakitani, M., Konno, Y., Yano, K., Shitara, K., Hosoi, S. & Satoh, M.
388
2006, "Comparison of cell lines for stable production of fucose-negative antibodies
with enhanced ADCC", Biotechnology and bioengineering, vol. 94, no. 4, pp. 680-
688.
Kang, H.J. & Park, H.J. 2009, "Novel molecular mechanism for actinomycin D activity
as an oncogenic promoter G-quadruplex binder", Biochemistry, vol. 48, no. 31, pp.
7392-7398.
Kang, S., Ren, D., Xiao, G., Daris, K., Buck, L., Enyenihi, A.A., Zubarev, R.,
Bondarenko, P.V. & Deshpande, R. 2013a, "Cell line profiling to improve
monoclonal antibody production", Biotechnology and bioengineering, 111(4), pp.
748-60.
Kang, S.M., Choi, J.W., Hong, S.H. & Lee, H.J. 2013b, "Up-Regulation of microRNA*
Strands by Their Target Transcripts", International journal of molecular sciences,
vol. 14, no. 7, pp. 13231-13240.
Kanitz, A. & Gerber, A.P. 2010, "Circuitry of mRNA regulation", Wiley interdisciplinary
reviews.Systems biology and medicine, vol. 2, no. 2, pp. 245-251.
Kantardjieff, A., Jacob, N.M., Yee, J.C., Epstein, E., Kok, Y.J., Philp, R., Betenbaugh,
M. & Hu, W.S. 2010, "Transcriptome and proteome analysis of Chinese hamster
ovary cells under low temperature and butyrate treatment", Journal of
Biotechnology, vol. 145, no. 2, pp. 143-159.
Karginov, F.V., Conaco, C., Xuan, Z., Schmidt, B.H., Parker, J.S., Mandel, G. & Hannon,
G.J. 2007, "A biochemical approach to identifying microRNA targets", Proceedings
of the National Academy of Sciences of the United States of America, vol. 104, no.
49, pp. 19291-19296.
Karreth, F.A., Tay, Y., Perna, D., Ala, U., Tan, S.M., Rust, A.G., DeNicola, G., Webster,
K.A., Weiss, D., Perez-Mancera, P.A., Krauthammer, M., Halaban, R., Provero, P.,
Adams, D.J., Tuveson, D.A. & Pandolfi, P.P. 2011, "In vivo identification of tumor-
suppressive PTEN ceRNAs in an oncogenic BRAF-induced mouse model of
melanoma", Cell, vol. 147, no. 2, pp. 382-395.
Kaufman, R.J. & Malhotra, J.D. 2014, "Calcium trafficking integrates endoplasmic
reticulum function with mitochondrial bioenergetics", Biochimica et biophysica
acta, 1843(10), pp. 2233-9.
Kawabe, Y., Makitsubo, H., Kameyama, Y., Huang, S., Ito, A. & Kamihira, M. 2012,
"Repeated integration of antibody genes into a pre-selected chromosomal locus of
CHO cells using an accumulative site-specific gene integration system",
Cytotechnology, vol. 64, no. 3, pp. 267-279.
Kertesz, M., Iovino, N., Unnerstall, U., Gaul, U. & Segal, E. 2007, "The role of site
accessibility in microRNA target recognition", Nature genetics, vol. 39, no. 10, pp.
1278-1284.
389
Ketting, R.F., Fischer, S.E., Bernstein, E., Sijen, T., Hannon, G.J. & Plasterk, R.H. 2001,
"Dicer functions in RNA interference and in synthesis of small RNA involved in
developmental timing in C. elegans", Genes & development, vol. 15, no. 20, pp.
2654-2659.
Khvorova, A., Reynolds, A. & Jayasena, S.D. 2003, "Functional siRNAs and miRNAs
exhibit strand bias", Cell, vol. 115, no. 2, pp. 209-216.
Kim do, Y., Chaudhry, M.A., Kennard, M.L., Jardon, M.A., Braasch, K., Dionne, B.,
Butler, M. & Piret, J.M. 2013, "Fed-batch CHO cell t-PA production and feed
glutamine replacement to reduce ammonia production", Biotechnology progress,
vol. 29, no. 1, pp. 165-175.
Kim, Y.G., Kim, J.Y. & Lee, G.M. 2009, "Effect of XIAP overexpression on sodium
butyrate-induced apoptosis in recombinant Chinese hamster ovary cells producing
erythropoietin", Journal of Biotechnology, vol. 144, no. 4, pp. 299-303.
Kim, H.R., Roe, J.S., Lee, J.E., Hwang, I.Y., Cho, E.J. & Youn, H.D. 2012, "A p53-
inducible microRNA-34a downregulates Ras signaling by targeting IMPDH",
Biochemical and biophysical research communications, vol. 418, no. 4, pp. 682-688.
Kim, H.S. & Lee, G.M. 2007, "Differences in optimal pH and temperature for cell growth
and antibody production between two Chinese hamster ovary clones derived from
the same parental clone", Journal of microbiology and biotechnology, vol. 17, no. 5,
pp. 712-720.
Kim, J.Y., Kim, Y.G., Han, Y.K., Choi, H.S., Kim, Y.H. & Lee, G.M. 2011, "Proteomic
understanding of intracellular responses of recombinant Chinese hamster ovary cells
cultivated in serum-free medium supplemented with hydrolysates", Applied
Microbiology and Biotechnology, vol. 89, no. 6, pp. 1917-1928.
Kim, J.Y., Kim, Y.G. & Lee, G.M. 2012, "CHO cells in biotechnology for production of
recombinant proteins: current state and further potential", Applied Microbiology and
Biotechnology, vol. 93, no. 3, pp. 917-930.
Kim, N.H., Kim, H.S., Kim, N.G., Lee, I., Choi, H.S., Li, X.Y., Kang, S.E., Cha, S.Y.,
Ryu, J.K., Na, J.M., Park, C., Kim, K., Lee, S., Gumbiner, B.M., Yook, J.I. & Weiss,
S.J. 2011, "p53 and microRNA-34 are suppressors of canonical Wnt signaling",
Science signaling, vol. 4, no. 197, pp. ra71.
Kim, N.S. & Lee, G.M. 2002, "Inhibition of sodium butyrate-induced apoptosis in
recombinant Chinese hamster ovary cells by constitutively expressing antisense
RNA of caspase-3", Biotechnology and bioengineering, vol. 78, no. 2, pp. 217-228.
Kim, S., Kim, D.H., Jung, W.H. & Koo, J.S. 2013, "Succinate dehydrogenase expression
in breast cancer", SpringerPlus, vol. 2, no. 1, pp. 299.
Kim, S.H. & Lee, G.M. 2007, "Down-regulation of lactate dehydrogenase-A by siRNAs
for reduced lactic acid formation of Chinese hamster ovary cells producing
390
thrombopoietin", Applied Microbiology and Biotechnology, vol. 74, no. 1, pp. 152-
159.
Kim, S.Y., Lee, S.M., Tak, J.K., Choi, K.S., Kwon, T.K. & Park, J.W. 2007, "Regulation
of singlet oxygen-induced apoptosis by cytosolic NADP+-dependent isocitrate
dehydrogenase", Molecular and cellular biochemistry, vol. 302, no. 1-2, pp. 27-34.
Kim, Y.G., Kim, J.Y., Mohan, C. & Lee, G.M. 2009a, "Effect of Bcl-xL overexpression
on apoptosis and autophagy in recombinant Chinese hamster ovary cells under
nutrient-deprived condition", Biotechnology and bioengineering, vol. 103, no. 4, pp.
757-766.
Kim, Y.K., Yu, J., Han, T.S., Park, S.Y., Namkoong, B., Kim, D.H., Hur, K., Yoo, M.W.,
Lee, H.J., Yang, H.K. & Kim, V.N. 2009b, "Functional links between clustered
microRNAs: suppression of cell-cycle inhibitors by microRNA clusters in gastric
cancer", Nucleic acids research, vol. 37, no. 5, pp. 1672-1681.
Kinoshita, T., Nohata, N., Yoshino, H., Hanazawa, T., Kikkawa, N., Fujimura, L.,
Chiyomaru, T., Kawakami, K., Enokida, H., Nakagawa, M., Okamoto, Y. & Seki,
N. 2012, "Tumor suppressive microRNA-375 regulates lactate dehydrogenase B in
maxillary sinus squamous cell carcinoma", International journal of oncology, vol.
40, no. 1, pp. 185-193.
Kitami, T., Logan, D.J., Negri, J., Hasaka, T., Tolliday, N.J., Carpenter, A.E.,
Spiegelman, B.M. & Mootha, V.K. 2012, "A chemical screen probing the
relationship between mitochondrial content and cell size", PloS one, vol. 7, no. 3, pp.
e33755.
Kitsera, N., Khobta, A. & Epe, B. 2007, "Destabilized green fluorescent protein detects
rapid removal of transcription blocks after genotoxic exposure", BioTechniques, vol.
43, no. 2, pp. 222-227.
Klanert, G., Jadhav, V., Chanoumidou, K., Grillari, J., Borth, N. & Hackl, M. 2013,
"Endogenous MicroRNA Clusters Outperform Chimeric Sequence Clusters in
Chinese Hamster Ovary Cells", Biotechnology journal, 9(4), pp. 538-44.
Kleppe, R., Martinez, A., Doskeland, S.O. & Haavik, J. 2011, "The 14-3-3 proteins in
regulation of cellular metabolism", Seminars in cell & developmental biology, vol.
22, no. 7, pp. 713-719.
Klionsky, D.J., Cuervo, A.M. & Seglen, P.O. 2007, "Methods for monitoring autophagy
from yeast to human", Autophagy, vol. 3, no. 3, pp. 181-206.
Kloosterman, W.P., Wienholds, E., Ketting, R.F. & Plasterk, R.H. 2004, "Substrate
requirements for let-7 function in the developing zebrafish embryo", Nucleic acids
research, vol. 32, no. 21, pp. 6284-6291.
Kluiver, J., Gibcus, J.H., Hettinga, C., Adema, A., Richter, M.K., Halsema, N., Slezak-
Prochazka, I., Ding, Y., Kroesen, B.J. & van den Berg, A. 2012a, "Rapid generation
of microRNA sponges for microRNA inhibition", PloS one, vol. 7, no. 1, pp. e29275.
391
Kluiver, J., Slezak-Prochazka, I., Smigielska-Czepiel, K., Halsema, N., Kroesen, B.J. &
van den Berg, A. 2012b, "Generation of miRNA sponge constructs", Methods (San
Diego, Calif.), 58(2), pp. 113-7.
Kobayashi, H., Tan, E.M. & Fleming, S.E. 2004, "Acetylation of histones associated with
the p21WAF1/CIP1 gene by butyrate is not sufficient for p21WAF1/CIP1 gene
transcription in human colorectal adenocarcinoma cells", International journal of
cancer.Journal international du cancer, vol. 109, no. 2, pp. 207-213.
Kondo, N., Toyama, T., Sugiura, H., Fujii, Y. & Yamashita, H. 2008, "miR-206
Expression is down-regulated in estrogen receptor alpha-positive human breast
cancer", Cancer research, vol. 68, no. 13, pp. 5004-5008.
Kong, K.Y., Owens, K.S., Rogers, J.H., Mullenix, J., Velu, C.S., Grimes, H.L. & Dahl,
R. 2010, "MIR-23A microRNA cluster inhibits B-cell development", Experimental
hematology, vol. 38, no. 8, pp. 629-640.e1.
Koscianska, E., Starega-Roslan, J. & Krzyzosiak, W.J. 2011, "The role of Dicer protein
partners in the processing of microRNA precursors", PloS one, vol. 6, no. 12, pp.
e28548.
Kou, T.C., Fan, L., Zhou, Y., Ye, Z.Y., Liu, X.P., Zhao, L. & Tan, W.S. 2011, "Detailed
understanding of enhanced specific productivity in Chinese hamster ovary cells at
low culture temperature", Journal of bioscience and bioengineering, vol. 111, no. 3,
pp. 365-369.
Kozomara, A. & Griffiths-Jones, S. 2011, "miRBase: integrating microRNA annotation
and deep-sequencing data", Nucleic acids research, vol. 39, no. Database issue, pp.
D152-7.
Krapp, S., Mimura, Y., Jefferis, R., Huber, R. & Sondermann, P. 2003, "Structural
analysis of human IgG-Fc glycoforms reveals a correlation between glycosylation
and structural integrity", Journal of Molecular Biology, vol. 325, no. 5, pp. 979-989.
Krek, A., Grun, D., Poy, M.N., Wolf, R., Rosenberg, L., Epstein, E.J., MacMenamin, P.,
da Piedade, I., Gunsalus, K.C., Stoffel, M. & Rajewsky, N. 2005, "Combinatorial
microRNA target predictions", Nature genetics, vol. 37, no. 5, pp. 495-500.
Krol, J., Loedige, I. & Filipowicz, W. 2010, "The widespread regulation of microRNA
biogenesis, function and decay", Nature reviews.Genetics, vol. 11, no. 9, pp. 597-
610.
Ku, S.C., Ng, D.T., Yap, M.G. & Chao, S.H. 2008, "Effects of overexpression of X-box
binding protein 1 on recombinant protein production in Chinese hamster ovary and
NS0 myeloma cells", Biotechnology and bioengineering, vol. 99, no. 1, pp. 155-164.
Kulshreshtha, R., Ferracin, M., Wojcik, S.E., Garzon, R., Alder, H., Agosto-Perez, F.J.,
Davuluri, R., Liu, C.G., Croce, C.M., Negrini, M., Calin, G.A. & Ivan, M. 2007, "A
microRNA signature of hypoxia", Molecular and cellular biology, vol. 27, no. 5, pp.
1859-1867.
392
Kumar, N., Gammell, P. & Clynes, M. 2007, "Proliferation control strategies to improve
productivity and survival during CHO based production culture: A summary of
recent methods employed and the effects of proliferation control in product secreting
CHO cell lines", Cytotechnology, vol. 53, no. 1-3, pp. 33-46.
Kumar, N., Gammell, P., Meleady, P., Henry, M. & Clynes, M. 2008, "Differential
protein expression following low temperature culture of suspension CHO-K1 cells",
BMC biotechnology, vol. 8, pp. 42-6750-8-42.
Kuystermans, D. & Al-Rubeai, M. 2009, "cMyc increases cell number through
uncoupling of cell division from cell size in CHO cells", BMC biotechnology, vol. 9,
pp. 76-6750-9-76.
Lagana, A., Russo, F., Sismeiro, C., Giugno, R., Pulvirenti, A. & Ferro, A. 2010,
"Variability in the incidence of miRNAs and genes in fragile sites and the role of
repeats and CpG islands in the distribution of genetic material", PloS one, vol. 5, no.
6, pp. e11166.
Lagos-Quintana, M., Rauhut, R., Lendeckel, W. & Tuschl, T. 2001, "Identification of
novel genes coding for small expressed RNAs", Science (New York, N.Y.), vol. 294,
no. 5543, pp. 853-858.
Lai, T., Yang, Y. & Ng, S.K. 2013a, "Advances in Mammalian cell line development
technologies for recombinant protein production", Pharmaceuticals (Basel,
Switzerland), vol. 6, no. 5, pp. 579-603.
Lai, T., Yang, Y. & Ng, S.K. 2013b, "Advances in Mammalian cell line development
technologies for recombinant protein production", Pharmaceuticals (Basel,
Switzerland), vol. 6, no. 5, pp. 579-603.
Lal, A., Navarro, F., Maher, C.A., Maliszewski, L.E., Yan, N., O'Day, E., Chowdhury,
D., Dykxhoorn, D.M., Tsai, P., Hofmann, O., Becker, K.G., Gorospe, M., Hide, W.
& Lieberman, J. 2009a, "miR-24 Inhibits cell proliferation by targeting E2F2, MYC,
and other cell-cycle genes via binding to "seedless" 3'UTR microRNA recognition
elements", Molecular cell, vol. 35, no. 5, pp. 610-625.
Lal, A., Navarro, F., Maher, C.A., Maliszewski, L.E., Yan, N., O'Day, E., Chowdhury,
D., Dykxhoorn, D.M., Tsai, P., Hofmann, O., Becker, K.G., Gorospe, M., Hide, W.
& Lieberman, J. 2009b, "miR-24 Inhibits cell proliferation by targeting E2F2, MYC,
and other cell-cycle genes via binding to "seedless" 3'UTR microRNA recognition
elements", Molecular cell, vol. 35, no. 5, pp. 610-625.
Lal, A., Thomas, M.P., Altschuler, G., Navarro, F., O'Day, E., Li, X.L., Concepcion, C.,
Han, Y.C., Thiery, J., Rajani, D.K., Deutsch, A., Hofmann, O., Ventura, A., Hide,
W. & Lieberman, J. 2011, "Capture of microRNA-bound mRNAs identifies the
tumor suppressor miR-34a as a regulator of growth factor signaling", PLoS genetics,
vol. 7, no. 11, pp. e1002363.
393
Larsen, F.J., Schiffer, T.A., Borniquel, S., Sahlin, K., Ekblom, B., Lundberg, J.O. &
Weitzberg, E. 2011, "Dietary inorganic nitrate improves mitochondrial efficiency in
humans", Cell metabolism, vol. 13, no. 2, pp. 149-159.
Larsen, S., Nielsen, J., Hansen, C.N., Nielsen, L.B., Wibrand, F., Stride, N., Schroder,
H.D., Boushel, R., Helge, J.W., Dela, F. & Hey-Mogensen, M. 2012, "Biomarkers
of mitochondrial content in skeletal muscle of healthy young human subjects", The
Journal of physiology, vol. 590, no. Pt 14, pp. 3349-3360.
Lau, N.C., Lim, L.P., Weinstein, E.G. & Bartel, D.P. 2001, "An abundant class of tiny
RNAs with probable regulatory roles in Caenorhabditis elegans", Science (New York,
N.Y.), vol. 294, no. 5543, pp. 858-862.
Laurila, E.M., Sandstrom, S., Rantanen, L.M., Autio, R. & Kallioniemi, A. 2012, "Both
inhibition and enhanced expression of miR-31 lead to reduced migration and
invasion of pancreatic cancer cells", Genes, chromosomes & cancer, vol. 51, no. 6,
pp. 557-568.
Le, A., Lane, A.N., Hamaker, M., Bose, S., Gouw, A., Barbi, J., Tsukamoto, T., Rojas,
C.J., Slusher, B.S., Zhang, H., Zimmerman, L.J., Liebler, D.C., Slebos, R.J.,
Lorkiewicz, P.K., Higashi, R.M., Fan, T.W. & Dang, C.V. 2012, "Glucose-
independent glutamine metabolism via TCA cycling for proliferation and survival in
B cells", Cell metabolism, vol. 15, no. 1, pp. 110-121.
Lebbink, R.J., Lowe, M., Chan, T., Khine, H., Wang, X. & McManus, M.T. 2011,
"Polymerase II promoter strength determines efficacy of microRNA adapted
shRNAs", PloS one, vol. 6, no. 10, pp. e26213.
Lee, H.C., Yang, C.W., Chen, C.Y. & Au, L.C. 2011, "Single point mutation of
microRNA may cause butterfly effect on alteration of global gene expression",
Biochemical and biophysical research communications, vol. 404, no. 4, pp. 1065-
1069.
Lee, J.S., Ha, T.K., Park, J.H. & Lee, G.M. 2013, "Anti-cell death engineering of CHO
cells: co-overexpression of Bcl-2 for apoptosis inhibition, Beclin-1 for autophagy
induction", Biotechnology and bioengineering, vol. 110, no. 8, pp. 2195-2207.
Lee, J.S. & Lee, G.M. 2012, "Effect of sodium butyrate on autophagy and apoptosis in
Chinese hamster ovary cells", Biotechnology progress, vol. 28, no. 2, pp. 349-357.
Lee, M.S., Kim, K.W., Kim, Y.H. & Lee, G.M. 2003, "Proteome analysis of antibody-
expressing CHO cells in response to hyperosmotic pressure", Biotechnology
progress, vol. 19, no. 6, pp. 1734-1741.
Lee, R., Feinbaum, R. & Ambros, V. 2004, "A short history of a short RNA", Cell, vol.
116, no. 2 Suppl, pp. S89-92, 1 p following S96.
Lee, R.C. & Ambros, V. 2001, "An extensive class of small RNAs in Caenorhabditis
elegans", Science (New York, N.Y.), vol. 294, no. 5543, pp. 862-864.
394
Lee, R.C., Feinbaum, R.L. & Ambros, V. 1993, "The C. elegans heterochronic gene lin-
4 encodes small RNAs with antisense complementarity to lin-14", Cell, vol. 75, no.
5, pp. 843-854.
Lee, S.J., Jiko, C., Yamashita, E. & Tsukihara, T. 2011, "Selective nuclear export
mechanism of small RNAs", Current opinion in structural biology, vol. 21, no. 1,
pp. 101-108.
Lee, S.M., Koh, H.J., Park, D.C., Song, B.J., Huh, T.L. & Park, J.W. 2002, "Cytosolic
NADP(+)-dependent isocitrate dehydrogenase status modulates oxidative damage to
cells", Free radical biology & medicine, vol. 32, no. 11, pp. 1185-1196.
Leivonen, S.K., Rokka, A., Ostling, P., Kohonen, P., Corthals, G.L., Kallioniemi, O. &
Perala, M. 2011, "Identification of miR-193b targets in breast cancer cells and
systems biological analysis of their functional impact", Molecular & cellular
proteomics: MCP, vol. 10, no. 7, pp. M110.005322.
Lempiainen, H. & Shore, D. 2009, "Growth control and ribosome biogenesis", Current
opinion in cell biology, vol. 21, no. 6, pp. 855-863.
Lennox, K.A. & Behlke, M.A. 2010, "A direct comparison of anti-microRNA
oligonucleotide potency", Pharmaceutical research, vol. 27, no. 9, pp. 1788-1799.
Levine, B. & Yuan, J. 2005, "Autophagy in cell death: an innocent convict?", The Journal
of clinical investigation, vol. 115, no. 10, pp. 2679-2688.
Lewis, B.P., Shih, I.H., Jones-Rhoades, M.W., Bartel, D.P. & Burge, C.B. 2003,
"Prediction of mammalian microRNA targets", Cell, vol. 115, no. 7, pp. 787-798.
Lewis, N.E., Liu, X., Li, Y., Nagarajan, H., Yerganian, G., O'Brien, E., Bordbar, A., Roth,
A.M., Rosenbloom, J., Bian, C., Xie, M., Chen, W., Li, N., Baycin-Hizal, D., Latif,
H., Forster, J., Betenbaugh, M.J., Famili, I., Xu, X., Wang, J. & Palsson, B.O. 2013,
"Genomic landscapes of Chinese hamster ovary cell lines as revealed by the
Cricetulus griseus draft genome", Nature biotechnology, vol. 31, no. 8, pp. 759-765.
Ley, D., Harraghy, N., Le Fourn, V., Bire, S., Girod, P.A., Regamey, A., Rouleux-Bonnin,
F., Bigot, Y. & Mermod, N. 2013, "MAR elements and transposons for improved
transgene integration and expression", PloS one, vol. 8, no. 4, pp. e62784.
Li, B., Sun, M., Gao, F., Liu, W., Yang, Y., Liu, H., Cheng, Y., Liu, C. & Cai, J. 2013a,
"Up-regulated expression of miR-23a/b targeted the pro-apoptotic Fas in radiation-
induced thymic lymphoma", Cellular physiology and biochemistry : international
journal of experimental cellular physiology, biochemistry, and pharmacology, vol.
32, no. 6, pp. 1729-1740.
Li, C., Rossomando, A., Wu, S.L. & Karger, B.L. 2013b, "Comparability analysis of anti-
CD20 commercial (rituximab) and RNAi-mediated fucosylated antibodies by two
LC-MS approaches", mAbs, vol. 5, no. 4, pp. 565-575.
395
Li, C., Wang, Y., Wang, S., Wu, B., Hao, J., Fan, H., Ju, Y., Ding, Y., Chen, L., Chu, X.,
Liu, W., Ye, X. & Meng, S. 2013c, "Hepatitis B virus mRNA-mediated miR-122
inhibition upregulates PTTG1-binding protein, which promotes hepatocellular
carcinoma tumor growth and cell invasion", Journal of virology, vol. 87, no. 4, pp.
2193-2205.
Li, F., Vijayasankaran, N., Shen, A.Y., Kiss, R. & Amanullah, A. 2010, "Cell culture
processes for monoclonal antibody production", mAbs, vol. 2, no. 5, pp. 466-479.
Li, F., Wang, Y., Zeller, K.I., Potter, J.J., Wonsey, D.R., O'Donnell, K.A., Kim, J.W.,
Yustein, J.T., Lee, L.A. & Dang, C.V. 2005, "Myc stimulates nuclearly encoded
mitochondrial genes and mitochondrial biogenesis", Molecular and cellular biology,
vol. 25, no. 14, pp. 6225-6234.
Li, H.Y., Zhang, Y., Cai, J.H. & Bian, H.L. 2013, "MicroRNA-451 inhibits growth of
human colorectal carcinoma cells via downregulation of Pi3k/Akt pathway", Asian
Pacific journal of cancer prevention: APJCP, vol. 14, no. 6, pp. 3631-3634.
Li, J., Menzel, C., Meier, D., Zhang, C., Dubel, S. & Jostock, T. 2007, "A comparative
study of different vector designs for the mammalian expression of recombinant IgG
antibodies", Journal of immunological methods, vol. 318, no. 1-2, pp. 113-124.
Li, S., Moffett, H.F., Lu, J., Werner, L., Zhang, H., Ritz, J., Neuberg, D., Wucherpfennig,
K.W., Brown, J.R. & Novina, C.D. 2011a, "MicroRNA expression profiling
identifies activated B cell status in chronic lymphocytic leukemia cells", PloS one,
vol. 6, no. 3, pp. e16956.
Li, W.Q., Chen, C., Xu, M.D., Guo, J., Li, Y.M., Xia, Q.M., Liu, H.M., He, J., Yu, H.Y.
& Zhu, L. 2011b, "The rno-miR-34 family is upregulated and targets ACSL1 in
dimethylnitrosamine-induced hepatic fibrosis in rats", The FEBS journal, vol. 278,
no. 9, pp. 1522-1532.
Li, X. 2014, "MiR-375, a microRNA related to diabetes", Gene, vol. 533, no. 1, pp. 1-4.
Li, X., Liu, X., Xu, W., Zhou, P., Gao, P., Jiang, S., Lobie, P.E. & Zhu, T. 2013, "c-MYC-
regulated miR-23a/24-2/27a cluster promotes mammary carcinoma cell invasion and
hepatic metastasis by targeting Sprouty2", The Journal of biological chemistry, vol.
288, no. 25, pp. 18121-18133.
Liew, J.C., Tan, W.S., Alitheen, N.B., Chan, E.S. & Tey, B.T. 2010, "Over-expression of
the X-linked inhibitor of apoptosis protein (XIAP) delays serum deprivation-induced
apoptosis in CHO-K1 cells", Journal of bioscience and bioengineering, vol. 110, no.
3, pp. 338-344.
Lim, Y., Seah, V.X., Mantalaris, A., Yap, M.G. & Wong, D.C. 2010, "Elucidating the
role of requiem in the growth and death of Chinese hamster ovary cells", Apoptosis
: An International Journal on Programmed Cell Death, vol. 15, no. 4, pp. 450-462.
396
Lin, N., Davis, A., Bahr, S., Borgschulte, T., Achtien, K. & Kayser, K. 2010, "Profiling
highly conserved microrna expression in recombinant IgG-producing and parental
chinese hamster ovary cells", Biotechnology progress, 27(4), pp. 1163-71.
Lin, Q., Gao, Z., Alarcon, R.M., Ye, J. & Yun, Z. 2009, "A role of miR-27 in the
regulation of adipogenesis", The FEBS journal, vol. 276, no. 8, pp. 2348-2358.
Little, G.H., Bai, Y., Williams, T. & Poizat, C. 2007, "Nuclear calcium/calmodulin-
dependent protein kinase IIdelta preferentially transmits signals to histone
deacetylase 4 in cardiac cells", The Journal of biological chemistry, vol. 282, no. 10,
pp. 7219-7231.
Little, P. 1993, "Genetics. Small and perfectly formed", Nature, vol. 366, no. 6452, pp.
204-205.
Liu, C., Chu, I. & Hwang, S. 2001, "Pentanoic acid, a novel protein synthesis stimulant
for Chinese Hamster Ovary (CHO) cells", Journal of bioscience and bioengineering,
vol. 91, no. 1, pp. 71-75.
Liu, C., Kelnar, K., Liu, B., Chen, X., Calhoun-Davis, T., Li, H., Patrawala, L., Yan, H.,
Jeter, C., Honorio, S., Wiggins, J.F., Bader, A.G., Fagin, R., Brown, D. & Tang, D.G.
2011, "The microRNA miR-34a inhibits prostate cancer stem cells and metastasis by
directly repressing CD44", Nature medicine, vol. 17, no. 2, pp. 211-215.
Liu, C.J., Tsai, M.M., Hung, P.S., Kao, S.Y., Liu, T.Y., Wu, K.J., Chiou, S.H., Lin, S.C.
& Chang, K.W. 2010, "miR-31 ablates expression of the HIF regulatory factor FIH
to activate the HIF pathway in head and neck carcinoma", Cancer research, vol. 70,
no. 4, pp. 1635-1644.
Liu, J., Carmell, M.A., Rivas, F.V., Marsden, C.G., Thomson, J.M., Song, J.J., Hammond,
S.M., Joshua-Tor, L. & Hannon, G.J. 2004, "Argonaute2 is the catalytic engine of
mammalian RNAi", Science (New York, N.Y.), vol. 305, no. 5689, pp. 1437-1441.
Liu, N., Zhang, J., Jiao, T., Li, Z., Peng, J., Cui, Z. & Ye, X. 2013, "Hepatitis B virus
inhibits apoptosis of hepatoma cells by sponging the MicroRNA 15a/16 cluster",
Journal of virology, vol. 87, no. 24, pp. 13370-13378.
Liu, W., Le, A., Hancock, C., Lane, A.N., Dang, C.V., Fan, T.W. & Phang, J.M. 2012,
"Reprogramming of proline and glutamine metabolism contributes to the
proliferative and metabolic responses regulated by oncogenic transcription factor c-
MYC", Proceedings of the National Academy of Sciences of the United States of
America, vol. 109, no. 23, pp. 8983-8988.
Liu, W., Zabirnyk, O., Wang, H., Shiao, Y.H., Nickerson, M.L., Khalil, S., Anderson,
L.M., Perantoni, A.O. & Phang, J.M. 2010a, "miR-23b targets proline oxidase, a
novel tumor suppressor protein in renal cancer", Oncogene, vol. 29, no. 35, pp. 4914-
4924.
Liu, X., Sempere, L.F., Ouyang, H., Memoli, V.A., Andrew, A.S., Luo, Y., Demidenko,
E., Korc, M., Shi, W., Preis, M., Dragnev, K.H., Li, H., Direnzo, J., Bak, M.,
397
Freemantle, S.J., Kauppinen, S. & Dmitrovsky, E. 2010b, "MicroRNA-31 functions
as an oncogenic microRNA in mouse and human lung cancer cells by repressing
specific tumor suppressors", The Journal of clinical investigation, vol. 120, no. 4,
pp. 1298-1309.
Liu, Y., Borchert, G.L., Donald, S.P., Diwan, B.A., Anver, M. & Phang, J.M. 2009,
"Proline oxidase functions as a mitochondrial tumor suppressor in human cancers",
Cancer research, vol. 69, no. 16, pp. 6414-6422.
Liu, Y., Borchert, G.L., Surazynski, A., Hu, C.A. & Phang, J.M. 2006, "Proline oxidase
activates both intrinsic and extrinsic pathways for apoptosis: the role of
ROS/superoxides, NFAT and MEK/ERK signaling", Oncogene, vol. 25, no. 41, pp.
5640-5647.
Lodygin, D., Tarasov, V., Epanchintsev, A., Berking, C., Knyazeva, T., Korner, H.,
Knyazev, P., Diebold, J. & Hermeking, H. 2008, "Inactivation of miR-34a by
aberrant CpG methylation in multiple types of cancer", Cell cycle (Georgetown,
Tex.), vol. 7, no. 16, pp. 2591-2600.
Loenarz, C. & Schofield, C.J. 2008, "Expanding chemical biology of 2-oxoglutarate
oxygenases", Nature chemical biology, vol. 4, no. 3, pp. 152-156.
Loh, W.P., Loo, B., Zhou, L., Zhang, P., Lee, D.Y., Yang, Y. & Lam, K.P. 2014,
"Overexpression of microRNAs enhances recombinant protein production in
Chinese hamster ovary cells", Biotechnology journal, 9(9), pp. 1140-51.
Lopert, P., Day, B.J. & Patel, M. 2012, "Thioredoxin reductase deficiency potentiates
oxidative stress, mitochondrial dysfunction and cell death in dopaminergic cells",
PloS one, vol. 7, no. 11, pp. e50683.
Lovis, P., Gattesco, S. & Regazzi, R. 2008, "Regulation of the expression of components
of the exocytotic machinery of insulin-secreting cells by microRNAs", Biological
chemistry, vol. 389, no. 3, pp. 305-312.
Lovis, P., Roggli, E., Laybutt, D.R., Gattesco, S., Yang, J.Y., Widmann, C.,
Abderrahmani, A. & Regazzi, R. 2008, "Alterations in microRNA expression
contribute to fatty acid-induced pancreatic beta-cell dysfunction", Diabetes, vol. 57,
no. 10, pp. 2728-2736.
Lupo, D., Vollmer, C., Deckers, M., Mick, D.U., Tews, I., Sinning, I. & Rehling, P. 2011,
"Mdm38 is a 14-3-3-like receptor and associates with the protein synthesis
machinery at the inner mitochondrial membrane", Traffic (Copenhagen, Denmark),
vol. 12, no. 10, pp. 1457-1466.
Lytle, J.R., Yario, T.A. & Steitz, J.A. 2007, "Target mRNAs are repressed as efficiently
by microRNA-binding sites in the 5' UTR as in the 3' UTR", Proceedings of the
National Academy of Sciences of the United States of America, vol. 104, no. 23, pp.
9667-9672.
398
Ma, J.B., Ye, K. & Patel, D.J. 2004, "Structural basis for overhang-specific small
interfering RNA recognition by the PAZ domain", Nature, vol. 429, no. 6989, pp.
318-322.
Ma, N., Ellet, J., Okediadi, C., Hermes, P., McCormick, E. & Casnocha, S. 2009, "A
single nutrient feed supports both chemically defined NS0 and CHO fed-batch
processes: Improved productivity and lactate metabolism", Biotechnology progress,
vol. 25, no. 5, pp. 1353-1363.
Majors, B.S., Arden, N., Oyler, G.A., Chiang, G.G., Pederson, N.E. & Betenbaugh, M.J.
2008, "E2F-1 overexpression increases viable cell density in batch cultures of
Chinese hamster ovary cells", Journal of Biotechnology, vol. 138, no. 3-4, pp. 103-
106.
Majors, B.S., Betenbaugh, M.J., Pederson, N.E. & Chiang, G.G. 2009, "Mcl-1
overexpression leads to higher viabilities and increased production of humanized
monoclonal antibody in Chinese hamster ovary cells", Biotechnology progress, vol.
25, no. 4, pp. 1161-1168.
Malhotra, J.D. & Kaufman, R.J. 2007, "Endoplasmic reticulum stress and oxidative
stress: a vicious cycle or a double-edged sword?", Antioxidants & redox signaling,
vol. 9, no. 12, pp. 2277-2293.
Mali, P., Esvelt, K.M. & Church, G.M. 2013, "Cas9 as a versatile tool for engineering
biology", Nature methods, vol. 10, no. 10, pp. 957-963.
Malphettes, L., Freyvert, Y., Chang, J., Liu, P.Q., Chan, E., Miller, J.C., Zhou, Z.,
Nguyen, T., Tsai, C., Snowden, A.W., Collingwood, T.N., Gregory, P.D. & Cost,
G.J. 2010, "Highly efficient deletion of FUT8 in CHO cell lines using zinc-finger
nucleases yields cells that produce completely nonfucosylated antibodies",
Biotechnology and bioengineering, vol. 106, no. 5, pp. 774-783.
Manganelli, F. & Ruggiero, L. 2013, "Clinical features of Pompe disease", Acta
myologica : myopathies and cardiomyopathies : official journal of the
Mediterranean Society of Myology / edited by the Gaetano Conte Academy for the
study of striated muscle diseases, vol. 32, no. 2, pp. 82-84.
Mansfield, J.H., Harfe, B.D., Nissen, R., Obenauer, J., Srineel, J., Chaudhuri, A., Farzan-
Kashani, R., Zuker, M., Pasquinelli, A.E., Ruvkun, G., Sharp, P.A., Tabin, C.J. &
McManus, M.T. 2004, "MicroRNA-responsive 'sensor' transgenes uncover Hox-like
and other developmentally regulated patterns of vertebrate microRNA expression",
Nature genetics, vol. 36, no. 10, pp. 1079-1083.
Marcinowski, L., Tanguy, M., Krmpotic, A., Radle, B., Lisnic, V.J., Tuddenham, L.,
Chane-Woon-Ming, B., Ruzsics, Z., Erhard, F., Benkartek, C., Babic, M., Zimmer,
R., Trgovcich, J., Koszinowski, U.H., Jonjic, S., Pfeffer, S. & Dolken, L. 2012,
"Degradation of cellular mir-27 by a novel, highly abundant viral transcript is
important for efficient virus replication in vivo", PLoS pathogens, vol. 8, no. 2, pp.
e1002510.
399
Mariani, M.R., Carpaneto, E.M., Ulivi, M., Allfrey, V.G. & Boffa, L.C. 2003,
"Correlation between butyrate-induced histone hyperacetylation turn-over and c-
myc expression", The Journal of steroid biochemistry and molecular biology, vol.
86, no. 2, pp. 167-171.
Marín-García, J. 2012, Mitochondria and Their Role in Cardiovascular Disease,
Springer.
Martinez, V.S., Dietmair, S., Quek, L.E., Hodson, M.P., Gray, P. & Nielsen, L.K. 2013,
"Flux balance analysis of CHO cells before and after a metabolic switch from lactate
production to consumption", Biotechnology and bioengineering, vol. 110, no. 2, pp.
660-666.
Martinez-Sanchez, A. & Murphy, C.L. 2013, "MicroRNA Target Identification—
Experimental Approaches", Biology, vol. 2, no. 1, pp. 189-205.
Matasci, M., Baldi, L., Hacker, D.L. & Wurm, F.M. 2011, "The PiggyBac transposon
enhances the frequency of CHO stable cell line generation and yields recombinant
lines with superior productivity and stability", Biotechnology and bioengineering,
vol. 108, no. 9, pp. 2141-2150.
Mazur, X., Fussenegger, M., Renner, W.A. & Bailey, J.E. 1998, "Higher productivity of
growth-arrested Chinese hamster ovary cells expressing the cyclin-dependent kinase
inhibitor p27", Biotechnology progress, vol. 14, no. 5, pp. 705-713.
Meents, H., Enenkel, B., Werner, R.G. & Fussenegger, M. 2002, "p27Kip1-mediated
controlled proliferation technology increases constitutive sICAM production in
CHO-DUKX adapted for growth in suspension and serum-free media",
Biotechnology and bioengineering, vol. 79, no. 6, pp. 619-627.
Meerbrey, K.L., Hu, G., Kessler, J.D., Roarty, K., Li, M.Z., Fang, J.E., Herschkowitz,
J.I., Burrows, A.E., Ciccia, A., Sun, T., Schmitt, E.M., Bernardi, R.J., Fu, X., Bland,
C.S., Cooper, T.A., Schiff, R., Rosen, J.M., Westbrook, T.F. & Elledge, S.J. 2011,
"The pINDUCER lentiviral toolkit for inducible RNA interference in vitro and in
vivo", Proceedings of the National Academy of Sciences of the United States of
America, vol. 108, no. 9, pp. 3665-3670.
Meister, G., Landthaler, M., Patkaniowska, A., Dorsett, Y., Teng, G. & Tuschl, T. 2004,
"Human Argonaute2 mediates RNA cleavage targeted by miRNAs and siRNAs",
Molecular cell, vol. 15, no. 2, pp. 185-197.
Meleady, P., Gallagher, M., Clarke, C., Henry, M., Sanchez, N., Barron, N. & Clynes, M.
2012a, "Impact of miR-7 over-expression on the proteome of Chinese hamster ovary
cells", Journal of Biotechnology, vol. 160, no. 3-4, pp. 251-262.
Meleady, P., Gallagher, M., Clarke, C., Henry, M., Sanchez, N., Barron, N. & Clynes, M.
2012b, "Impact of miR-7 over-expression on the proteome of Chinese hamster ovary
cells", Journal of Biotechnology, vol. 160, no. 3-4, pp. 251-262.
400
Meleady, P., Hoffrogge, R., Henry, M., Rupp, O., Bort, J.H., Clarke, C., Brinkrolf, K.,
Kelly, S., Muller, B., Doolan, P., Hackl, M., Beckmann, T.F., Noll, T., Grillari, J.,
Barron, N., Puhler, A., Clynes, M. & Borth, N. 2012c, "Utilization and evaluation of
CHO-specific sequence databases for mass spectrometry based proteomics",
Biotechnology and bioengineering, vol. 109, no. 6, pp. 1386-1394.
Memczak, S., Jens, M., Elefsinioti, A., Torti, F., Krueger, J., Rybak, A., Maier, L.,
Mackowiak, S.D., Gregersen, L.H., Munschauer, M., Loewer, A., Ziebold, U.,
Landthaler, M., Kocks, C., le Noble, F. & Rajewsky, N. 2013, "Circular RNAs are a
large class of animal RNAs with regulatory potency", Nature, vol. 495, no. 7441, pp.
333-338.
Mertens-Talcott, S.U., Chintharlapalli, S., Li, X. & Safe, S. 2007, "The oncogenic
microRNA-27a targets genes that regulate specificity protein transcription factors
and the G2-M checkpoint in MDA-MB-231 breast cancer cells", Cancer research,
vol. 67, no. 22, pp. 11001-11011.
Mimura, Y., Church, S., Ghirlando, R., Ashton, P.R., Dong, S., Goodall, M., Lund, J. &
Jefferis, R. 2000, "The influence of glycosylation on the thermal stability and
effector function expression of human IgG1-Fc: properties of a series of truncated
glycoforms", Molecular immunology, vol. 37, no. 12-13, pp. 697-706.
Mishima, Y., Fukao, A., Kishimoto, T., Sakamoto, H., Fujiwara, T. & Inoue, K. 2012,
"Translational inhibition by deadenylation-independent mechanisms is central to
microRNA-mediated silencing in zebrafish", Proceedings of the National Academy
of Sciences of the United States of America, vol. 109, no. 4, pp. 1104-1109.
Mittermayr, S., Bones, J., Doherty, M., Guttman, A. & Rudd, P.M. 2011, "Multiplexed
analytical glycomics: rapid and confident IgG N-glycan structural elucidation",
Journal of proteome research, vol. 10, no. 8, pp. 3820-3829.
Mizuochi, T., Taniguchi, T., Shimizu, A. & Kobata, A. 1982, "Structural and numerical
variations of the carbohydrate moiety of immunoglobulin G", Journal of
immunology (Baltimore, Md.: 1950), vol. 129, no. 5, pp. 2016-2020.
Mizushima, N., Yoshimori, T. & Levine, B. 2010, "Methods in mammalian autophagy
research", Cell, vol. 140, no. 3, pp. 313-326.
Mohan, C., Kim, Y.G., Koo, J. & Lee, G.M. 2008a, "Assessment of cell engineering
strategies for improved therapeutic protein production in CHO cells", Biotechnology
journal, vol. 3, no. 5, pp. 624-630.
Mohan, C., Kim, Y.G., Koo, J. & Lee, G.M. 2008b, "Assessment of cell engineering
strategies for improved therapeutic protein production in CHO cells", Biotechnology
journal, vol. 3, no. 5, pp. 624-630.
Mohan, C., Park, S.H., Chung, J.Y. & Lee, G.M. 2007, "Effect of doxycycline-regulated
protein disulfide isomerase expression on the specific productivity of recombinant
CHO cells: thrombopoietin and antibody", Biotechnology and bioengineering, vol.
98, no. 3, pp. 611-615.
401
Mori, K., Kuni-Kamochi, R., Yamane-Ohnuki, N., Wakitani, M., Yamano, K., Imai, H.,
Kanda, Y., Niwa, R., Iida, S., Uchida, K., Shitara, K. & Satoh, M. 2004, "Engineering
Chinese hamster ovary cells to maximize effector function of produced antibodies
using FUT8 siRNA", Biotechnology and bioengineering, vol. 88, no. 7, pp. 901-908.
Morin, R.D., O'Connor, M.D., Griffith, M., Kuchenbauer, F., Delaney, A., Prabhu, A.L.,
Zhao, Y., McDonald, H., Zeng, T., Hirst, M., Eaves, C.J. & Marra, M.A. 2008,
"Application of massively parallel sequencing to microRNA profiling and discovery
in human embryonic stem cells", Genome research, vol. 18, no. 4, pp. 610-621.
Moss, E.G., Lee, R.C. & Ambros, V. 1997, "The cold shock domain protein LIN-28
controls developmental timing in C. elegans and is regulated by the lin-4 RNA", Cell,
vol. 88, no. 5, pp. 637-646.
Mukherji, S., Ebert, M.S., Zheng, G.X., Tsang, J.S., Sharp, P.A. & van Oudenaarden, A.
2011, "MicroRNAs can generate thresholds in target gene expression", Nature
genetics, vol. 43, no. 9, pp. 854-859.
Muller, D., Katinger, H. & Grillari, J. 2008, "MicroRNAs as targets for engineering of
CHO cell factories", Trends in biotechnology, vol. 26, no. 7, pp. 359-365.
Mullokandov, G., Baccarini, A., Ruzo, A., Jayaprakash, A.D., Tung, N., Israelow, B.,
Evans, M.J., Sachidanandam, R. & Brown, B.D. 2012, "High-throughput assessment
of microRNA activity and function using microRNA sensor and decoy libraries",
Nature methods, vol. 9, no. 8, pp. 840-846.
Nam, S., Kim, B., Shin, S. & Lee, S. 2008, "miRGator: an integrated system for functional
annotation of microRNAs", Nucleic acids research, vol. 36, no. Database issue, pp.
D159-64.
Navarro, F., Gutman, D., Meire, E., Caceres, M., Rigoutsos, I., Bentwich, Z. &
Lieberman, J. 2009, "miR-34a contributes to megakaryocytic differentiation of K562
cells independently of p53", Blood, vol. 114, no. 10, pp. 2181-2192.
Neermann, J. & Wagner, R. 1996, "Comparative analysis of glucose and glutamine
metabolism in transformed mammalian cell lines, insect and primary liver cells",
Journal of cellular physiology, vol. 166, no. 1, pp. 152-169.
Niederer, F., Trenkmann, M., Ospelt, C., Karouzakis, E., Neidhart, M., Stanczyk, J.,
Kolling, C., Gay, R.E., Detmar, M., Gay, S., Jungel, A. & Kyburz, D. 2012, "Down-
regulation of microRNA-34a* in rheumatoid arthritis synovial fibroblasts promotes
apoptosis resistance", Arthritis and Rheumatism, vol. 64, no. 6, pp. 1771-1779.
Nimmerjahn, F. & Ravetch, J.V. 2008, "Fcgamma receptors as regulators of immune
responses", Nature reviews.Immunology, vol. 8, no. 1, pp. 34-47.
Nissom, P.M., Sanny, A., Kok, Y.J., Hiang, Y.T., Chuah, S.H., Shing, T.K., Lee, Y.Y.,
Wong, K.T., Hu, W.S., Sim, M.Y. & Philp, R. 2006, "Transcriptome and proteome
profiling to understanding the biology of high productivity CHO cells", Molecular
biotechnology, vol. 34, no. 2, pp. 125-140.
402
Norling, L.L., Colca, J.R., Kelly, P.T., McDaniel, M.L. & Landt, M. 1994, "Activation
of calcium and calmodulin dependent protein kinase II during stimulation of insulin
secretion", Cell calcium, vol. 16, no. 2, pp. 137-150.
Obad, S., dos Santos, C.O., Petri, A., Heidenblad, M., Broom, O., Ruse, C., Fu, C.,
Lindow, M., Stenvang, J., Straarup, E.M., Hansen, H.F., Koch, T., Pappin, D.,
Hannon, G.J. & Kauppinen, S. 2011, "Silencing of microRNA families by seed-
targeting tiny LNAs", Nature genetics, vol. 43, no. 4, pp. 371-378.
O'Donnell, K.A., Wentzel, E.A., Zeller, K.I., Dang, C.V. & Mendell, J.T. 2005a, "c-Myc-
regulated microRNAs modulate E2F1 expression", Nature, vol. 435, no. 7043, pp.
839-843.
O'Donnell, K.A., Wentzel, E.A., Zeller, K.I., Dang, C.V. & Mendell, J.T. 2005b, "c-Myc-
regulated microRNAs modulate E2F1 expression", Nature, vol. 435, no. 7043, pp.
839-843.
Oguchi, S., Saito, H., Tsukahara, M. & Tsumura, H. 2006, "pH Condition in temperature
shift cultivation enhances cell longevity and specific hMab productivity in CHO
culture", Cytotechnology, vol. 52, no. 3, pp. 199-207.
Ohya, T., Hayashi, T., Kiyama, E., Nishii, H., Miki, H., Kobayashi, K., Honda, K.,
Omasa, T. & Ohtake, H. 2008a, "Improved production of recombinant human
antithrombin III in Chinese hamster ovary cells by ATF4 overexpression",
Biotechnology and bioengineering, vol. 100, no. 2, pp. 317-324.
Ohya, T., Hayashi, T., Kiyama, E., Nishii, H., Miki, H., Kobayashi, K., Honda, K.,
Omasa, T. & Ohtake, H. 2008b, "Improved production of recombinant human
antithrombin III in Chinese hamster ovary cells by ATF4 overexpression",
Biotechnology and bioengineering, vol. 100, no. 2, pp. 317-324.
Okamura, K., Hagen, J.W., Duan, H., Tyler, D.M. & Lai, E.C. 2007, "The mirtron
pathway generates microRNA-class regulatory RNAs in Drosophila", Cell, vol. 130,
no. 1, pp. 89-100.
Okamura, K., Phillips, M.D., Tyler, D.M., Duan, H., Chou, Y.T. & Lai, E.C. 2008, "The
regulatory activity of microRNA* species has substantial influence on microRNA
and 3' UTR evolution", Nature structural & molecular biology, vol. 15, no. 4, pp.
354-363.
Olena, A.F. & Patton, J.G. 2010, "Genomic organization of microRNAs", Journal of
cellular physiology, vol. 222, no. 3, pp. 540-545.
Olive, V., Jiang, I. & He, L. 2010a, "mir-17-92, a cluster of miRNAs in the midst of the
cancer network", The international journal of biochemistry & cell biology, vol. 42,
no. 8, pp. 1348-1354.
Olive, V., Jiang, I. & He, L. 2010b, "mir-17-92, a cluster of miRNAs in the midst of the
cancer network", The international journal of biochemistry & cell biology, vol. 42,
no. 8, pp. 1348-1354.
403
Olive, V., Li, Q. & He, L. 2013, "Mir-17-92: a Polycistronic Oncomir with Pleiotropic
Functions", Immunological reviews, vol. 253, no. 1, pp. 158-166.
Omasa, T., Takami, T., Ohya, T., Kiyama, E., Hayashi, T., Nishii, H., Miki, H.,
Kobayashi, K., Honda, K. & Ohtake, H. 2008, "Overexpression of GADD34
enhances production of recombinant human antithrombin III in Chinese hamster
ovary cells", Journal of bioscience and bioengineering, vol. 106, no. 6, pp. 568-573.
Orom, U.A. & Lund, A.H. 2007, "Isolation of microRNA targets using biotinylated
synthetic microRNAs", Methods (San Diego, Calif.), vol. 43, no. 2, pp. 162-165.
Orom, U.A., Nielsen, F.C. & Lund, A.H. 2008, "MicroRNA-10a binds the 5'UTR of
ribosomal protein mRNAs and enhances their translation", Molecular cell, vol. 30,
no. 4, pp. 460-471.
Osellame, L.D., Blacker, T.S. & Duchen, M.R. 2012, "Cellular and molecular
mechanisms of mitochondrial function", Best practice & research.Clinical
endocrinology & metabolism, vol. 26, no. 6, pp. 711-723.
Osterhoff, M., Mohlig, M., Schwanstecher, M., Seufert, J., Ortmann, J., Schatz, H. &
Pfeiffer, A.F. 2003, "Ca2+/calmodulin-dependent protein kinase II delta2 regulates
gene expression of insulin in INS-1 rat insulinoma cells", Cell calcium, vol. 33, no.
3, pp. 175-184.
Ota, A., Tagawa, H., Karnan, S., Tsuzuki, S., Karpas, A., Kira, S., Yoshida, Y. & Seto,
M. 2004a, "Identification and characterization of a novel gene, C13orf25, as a target
for 13q31-q32 amplification in malignant lymphoma", Cancer research, vol. 64, no.
9, pp. 3087-3095.
Ota, A., Tagawa, H., Karnan, S., Tsuzuki, S., Karpas, A., Kira, S., Yoshida, Y. & Seto,
M. 2004b, "Identification and characterization of a novel gene, C13orf25, as a target
for 13q31-q32 amplification in malignant lymphoma", Cancer research, vol. 64, no.
9, pp. 3087-3095.
Ouyang, Y.B., Lu, Y., Yue, S. & Giffard, R.G. 2012, "miR-181 targets multiple Bcl-2
family members and influences apoptosis and mitochondrial function in astrocytes",
Mitochondrion, vol. 12, no. 2, pp. 213-219.
Owen, O.E., Kalhan, S.C. & Hanson, R.W. 2002, "The key role of anaplerosis and
cataplerosis for citric acid cycle function", The Journal of biological chemistry, vol.
277, no. 34, pp. 30409-30412.
Pan, S., Ryu, S.Y. & Sheu, S.S. 2011, "Distinctive characteristics and functions of
multiple mitochondrial Ca2+ influx mechanisms", Science China.Life sciences, vol.
54, no. 8, pp. 763-769.
Panzitt, K., Tschernatsch, M.M., Guelly, C., Moustafa, T., Stradner, M., Strohmaier,
H.M., Buck, C.R., Denk, H., Schroeder, R., Trauner, M. & Zatloukal, K. 2007,
"Characterization of HULC, a novel gene with striking up-regulation in
404
hepatocellular carcinoma, as noncoding RNA", Gastroenterology, vol. 132, no. 1,
pp. 330-342.
Park, C.Y., Choi, Y.S. & McManus, M.T. 2010, "Analysis of microRNA knockouts in
mice", Human molecular genetics, vol. 19, no. R2, pp. R169-75.
Parlato, S., Bruni, R., Fragapane, P., Salerno, D., Marcantonio, C., Borghi, P., Tataseo,
P., Ciccaglione, A.R., Presutti, C., Romagnoli, G., Bozzoni, I., Belardelli, F. &
Gabriele, L. 2013, "IFN-alpha regulates Blimp-1 expression via miR-23a and miR-
125b in both monocytes-derived DC and pDC", PloS one, vol. 8, no. 8, pp. e72833.
Pasquinelli, A.E. 2012, "MicroRNAs and their targets: recognition, regulation and an
emerging reciprocal relationship", Nature reviews.Genetics, vol. 13, no. 4, pp. 271-
282.
Pasquinelli, A.E., Reinhart, B.J., Slack, F., Martindale, M.Q., Kuroda, M.I., Maller, B.,
Hayward, D.C., Ball, E.E., Degnan, B., Muller, P., Spring, J., Srinivasan, A.,
Fishman, M., Finnerty, J., Corbo, J., Levine, M., Leahy, P., Davidson, E. & Ruvkun,
G. 2000, "Conservation of the sequence and temporal expression of let-7
heterochronic regulatory RNA", Nature, vol. 408, no. 6808, pp. 86-89.
Pattingre, S., Tassa, A., Qu, X., Garuti, R., Liang, X.H., Mizushima, N., Packer, M.,
Schneider, M.D. & Levine, B. 2005, "Bcl-2 antiapoptotic proteins inhibit Beclin 1-
dependent autophagy", Cell, vol. 122, no. 6, pp. 927-939.
Pekarsky, Y., Santanam, U., Cimmino, A., Palamarchuk, A., Efanov, A., Maximov, V.,
Volinia, S., Alder, H., Liu, C.G., Rassenti, L., Calin, G.A., Hagan, J.P., Kipps, T. &
Croce, C.M. 2006, "Tcl1 expression in chronic lymphocytic leukemia is regulated
by miR-29 and miR-181", Cancer research, vol. 66, no. 24, pp. 11590-11593.
Pelicano, H., Martin, D.S., Xu, R.H. & Huang, P. 2006a, "Glycolysis inhibition for
anticancer treatment", Oncogene, vol. 25, no. 34, pp. 4633-4646.
Pelicano, H., Xu, R.H., Du, M., Feng, L., Sasaki, R., Carew, J.S., Hu, Y., Ramdas, L.,
Hu, L., Keating, M.J., Zhang, W., Plunkett, W. & Huang, P. 2006b, "Mitochondrial
respiration defects in cancer cells cause activation of Akt survival pathway through
a redox-mediated mechanism", The Journal of cell biology, vol. 175, no. 6, pp. 913-
923.
Peng, R.W., Abellan, E. & Fussenegger, M. 2011, "Differential effect of exocytic
SNAREs on the production of recombinant proteins in mammalian cells",
Biotechnology and bioengineering, vol. 108, no. 3, pp. 611-620.
Peng, R.W. & Fussenegger, M. 2009, "Molecular engineering of exocytic vesicle traffic
enhances the productivity of Chinese hamster ovary cells", Biotechnology and
bioengineering, vol. 102, no. 4, pp. 1170-1181.
Perez-Mato, M., Ramos-Cabrer, P., Sobrino, T., Blanco, M., Ruban, A., Mirelman, D.,
Menendez, P., Castillo, J. & Campos, F. 2014, "Human recombinant glutamate
405
oxaloacetate transaminase 1 (GOT1) supplemented with oxaloacetate induces a
protective effect after cerebral ischemia", Cell death & disease, vol. 5, pp. e992.
Piao, L., Li, Y., Kim, S.J., Byun, H.S., Huang, S.M., Hwang, S.K., Yang, K.J., Park, K.A.,
Won, M., Hong, J., Hur, G.M., Seok, J.H., Shong, M., Cho, M.H., Brazil, D.P.,
Hemmings, B.A. & Park, J. 2009, "Association of LETM1 and MRPL36 contributes
to the regulation of mitochondrial ATP production and necrotic cell death", Cancer
research, vol. 69, no. 8, pp. 3397-3404.
Pichinuk, E., Broday, L. & Wreschner, D.H. 2011, "Endogenous RNA cleavages at the
ribosomal SRL site likely reflect miRNA (miR) mediated translational suppression",
Biochemical and biophysical research communications, vol. 414, no. 4, pp. 706-711.
Pichler, J., Galosy, S., Mott, J. & Borth, N. 2011, "Selection of CHO host cell subclones
with increased specific antibody production rates by repeated cycles of transient
transfection and cell sorting", Biotechnology and bioengineering, vol. 108, no. 2, pp.
386-394.
Pickering, A.M. & Davies, K.J. 2012, "Degradation of damaged proteins: the main
function of the 20S proteasome", Progress in molecular biology and translational
science, vol. 109, pp. 227-248.
Pilbrough, W., Munro, T.P. & Gray, P. 2009, "Intraclonal protein expression
heterogeneity in recombinant CHO cells", PloS one, vol. 4, no. 12, pp. e8432.
Pitceathly, R.D., Rahman, S., Wedatilake, Y., Polke, J.M., Cirak, S., Foley, A.R., Sailer,
A., Hurles, M.E., Stalker, J., Hargreaves, I., Woodward, C.E., Sweeney, M.G.,
Muntoni, F., Houlden, H., Taanman, J.W., Hanna, M.G. & UK10K Consortium
2013, "NDUFA4 mutations underlie dysfunction of a cytochrome c oxidase subunit
linked to human neurological disease", Cell reports, vol. 3, no. 6, pp. 1795-1805.
Poburko, D. & Demaurex, N. 2012, "Regulation of the mitochondrial proton gradient by
cytosolic Ca(2)(+) signals", Pflugers Archiv : European journal of physiology, vol.
464, no. 1, pp. 19-26.
Poliseno, L., Salmena, L., Zhang, J., Carver, B., Haveman, W.J. & Pandolfi, P.P. 2010,
"A coding-independent function of gene and pseudogene mRNAs regulates tumour
biology", Nature, vol. 465, no. 7301, pp. 1033-1038.
Polyak, K., Xia, Y., Zweier, J.L., Kinzler, K.W. & Vogelstein, B. 1997, "A model for
p53-induced apoptosis", Nature, vol. 389, no. 6648, pp. 300-305.
Pontiller, J., Gross, S., Thaisuchat, H., Hesse, F. & Ernst, W. 2008, "Identification of
CHO endogenous promoter elements based on a genomic library approach",
Molecular biotechnology, vol. 39, no. 2, pp. 135-139.
Ponting, C.P., Oliver, P.L. & Reik, W. 2009, "Evolution and functions of long noncoding
RNAs", Cell, vol. 136, no. 4, pp. 629-641.
406
Porter, A.J., Racher, A.J., Preziosi, R. & Dickson, A.J. 2010, "Strategies for selecting
recombinant CHO cell lines for cGMP manufacturing: improving the efficiency of
cell line generation", Biotechnology progress, vol. 26, no. 5, pp. 1455-1464.
Poy, M.N., Eliasson, L., Krutzfeldt, J., Kuwajima, S., Ma, X., Macdonald, P.E., Pfeffer,
S., Tuschl, T., Rajewsky, N., Rorsman, P. & Stoffel, M. 2004, "A pancreatic islet-
specific microRNA regulates insulin secretion", Nature, vol. 432, no. 7014, pp. 226-
230.
Pybus, L.P., Dean, G., West, N.R., Smith, A., Daramola, O., Field, R., Wilkinson, S.J. &
James, D.C. 2013, "Model-directed engineering of "difficult-to-express" monoclonal
antibody production by Chinese hamster ovary cells", Biotechnology and
bioengineering, 111(2), pp. 372-85.
Qiu, X., Friedman, J.M. & Liang, G. 2011, "Creating a flexible multiple microRNA
expression vector by linking precursor microRNAs", Biochemical and biophysical
research communications, vol. 411, no. 2, pp. 276-280.
Qu, B., Han, X., Tang, Y. & Shen, N. 2012, "A novel vector-based method for exclusive
overexpression of star-form microRNAs", PloS one, vol. 7, no. 7, pp. e41504.
Quan, H., Fan, G. & Wang, C.C. 1995, "Independence of the chaperone activity of protein
disulfide isomerase from its thioredoxin-like active site", The Journal of biological
chemistry, vol. 270, no. 29, pp. 17078-17080.
Raben, N., Wong, A., Ralston, E. & Myerowitz, R. 2012, "Autophagy and mitochondria
in Pompe disease: nothing is so new as what has long been forgotten", American
journal of medical genetics.Part C, Seminars in medical genetics, vol. 160C, no. 1,
pp. 13-21.
Rahimpour, A., Vaziri, B., Moazzami, R., Nematollahi, L., Barkhordari, F., Kokabee, L.,
Adeli, A. & Mahboudi, F. 2013, "Engineering the Cellular Protein Secretory
Pathway for Enhancement of Recombinant Tissue Plasminogen Activator
Expression in Chinese Hamster Ovary cells: Effects of CERT and XBP1s Genes",
Journal of microbiology and biotechnology, 23(8), pp. 1116-22.
Raju, T.S. 2008, "Terminal sugars of Fc glycans influence antibody effector functions of
IgGs", Current opinion in immunology, vol. 20, no. 4, pp. 471-478.
Raza, H. 2011, "Dual localization of glutathione S-transferase in the cytosol and
mitochondria: implications in oxidative stress, toxicity and disease", The FEBS
journal, vol. 278, no. 22, pp. 4243-4251.
Raza, H., Robin, M.A., Fang, J.K. & Avadhani, N.G. 2002, "Multiple isoforms of
mitochondrial glutathione S-transferases and their differential induction under
oxidative stress", The Biochemical journal, vol. 366, no. Pt 1, pp. 45-55.
Reichert, J.M. 2013, "Which are the antibodies to watch in 2013?", mAbs, vol. 5, no. 1,
pp. 1-4.
407
Reichert, J.M. 2010, "Antibodies to watch in 2010", mAbs, vol. 2, no. 1, pp. 84-100.
Reinhart, B.J., Slack, F.J., Basson, M., Pasquinelli, A.E., Bettinger, J.C., Rougvie, A.E.,
Horvitz, H.R. & Ruvkun, G. 2000, "The 21-nucleotide let-7 RNA regulates
developmental timing in Caenorhabditis elegans", Nature, vol. 403, no. 6772, pp.
901-906.
Reinheckel, T., Sitte, N., Ullrich, O., Kuckelkorn, U., Davies, K.J. & Grune, T. 1998,
"Comparative resistance of the 20S and 26S proteasome to oxidative stress", The
Biochemical journal, vol. 335 ( Pt 3), no. Pt 3, pp. 637-642.
Rhee, S.G., Woo, H.A., Kil, I.S. & Bae, S.H. 2012, "Peroxiredoxin functions as a
peroxidase and a regulator and sensor of local peroxides", The Journal of biological
chemistry, vol. 287, no. 7, pp. 4403-4410.
Rizzuto, R. & Pozzan, T. 2006, "Microdomains of intracellular Ca2+: molecular
determinants and functional consequences", Physiological Reviews, vol. 86, no. 1,
pp. 369-408.
Robertson, K.D. & Jones, P.A. 2000, "DNA methylation: past, present and future
directions", Carcinogenesis, vol. 21, no. 3, pp. 461-467.
Rolfe, D.F., Newman, J.M., Buckingham, J.A., Clark, M.G. & Brand, M.D. 1999,
"Contribution of mitochondrial proton leak to respiration rate in working skeletal
muscle and liver and to SMR", The American Journal of Physiology, vol. 276, no. 3
Pt 1, pp. C692-9.
Ronchetti, D., Lionetti, M., Mosca, L., Agnelli, L., Andronache, A., Fabris, S., Deliliers,
G.L. & Neri, A. 2008, "An integrative genomic approach reveals coordinated
expression of intronic miR-335, miR-342, and miR-561 with deregulated host genes
in multiple myeloma", BMC medical genomics, vol. 1, pp. 37.
Ronnebaum, S.M., Ilkayeva, O., Burgess, S.C., Joseph, J.W., Lu, D., Stevens, R.D.,
Becker, T.C., Sherry, A.D., Newgard, C.B. & Jensen, M.V. 2006, "A pyruvate
cycling pathway involving cytosolic NADP-dependent isocitrate dehydrogenase
regulates glucose-stimulated insulin secretion", The Journal of biological chemistry,
vol. 281, no. 41, pp. 30593-30602.
Roopenian, D.C. & Akilesh, S. 2007, "FcRn: the neonatal Fc receptor comes of age",
Nature reviews.Immunology, vol. 7, no. 9, pp. 715-725.
Rose, N.R., McDonough, M.A., King, O.N., Kawamura, A. & Schofield, C.J. 2011,
"Inhibition of 2-oxoglutarate dependent oxygenases", Chemical Society Reviews,
vol. 40, no. 8, pp. 4364-4397.
Rossi, J.J. 2008, "Expression strategies for short hairpin RNA interference triggers",
Human Gene Therapy, vol. 19, no. 4, pp. 313-317.
Ruepp, A., Kowarsch, A., Schmidl, D., Buggenthin, F., Brauner, B., Dunger, I., Fobo, G.,
Frishman, G., Montrone, C. & Theis, F.J. 2010, "PhenomiR: a knowledgebase for
408
microRNA expression in diseases and biological processes", Genome biology, vol.
11, no. 1, pp. R6.
Running Deer, J. & Allison, D.S. 2004, "High-level expression of proteins in mammalian
cells using transcription regulatory sequences from the Chinese hamster EF-1alpha
gene", Biotechnology progress, vol. 20, no. 3, pp. 880-889.
Ruvkun, G. & Giusto, J. 1989, "The Caenorhabditis elegans heterochronic gene lin-14
encodes a nuclear protein that forms a temporal developmental switch", Nature, vol.
338, no. 6213, pp. 313-319.
Ryll, T., Dutina, G., Reyes, A., Gunson, J., Krummen, L. & Etcheverry, T. 2000,
"Performance of small-scale CHO perfusion cultures using an acoustic cell filtration
device for cell retention: characterization of separation efficiency and impact of
perfusion on product quality", Biotechnology and bioengineering, vol. 69, no. 4, pp.
440-449.
Saini, H.K., Griffiths-Jones, S. & Enright, A.J. 2007, "Genomic analysis of human
microRNA transcripts", Proceedings of the National Academy of Sciences of the
United States of America, vol. 104, no. 45, pp. 17719-17724.
Salmena, L., Poliseno, L., Tay, Y., Kats, L. & Pandolfi, P.P. 2011, "A ceRNA hypothesis:
the Rosetta Stone of a hidden RNA language?", Cell, vol. 146, no. 3, pp. 353-358.
Sanchez, N., Gallagher, M., Lao, N., Gallagher, C., Clarke, C., Doolan, P., Aherne, S.,
Blanco, A., Meleady, P., Clynes, M. & Barron, N. 2013a, "MiR-7 Triggers Cell
Cycle Arrest at the G1/S Transition by Targeting Multiple Genes Including Skp2 and
Psme3", PloS one, vol. 8, no. 6, pp. e65671.
Sanchez, N., Kelly, P., Gallagher, C., Lao, N.T., Clarke, C., Clynes, M. & Barron, N.
2013b, "CHO cell culture longevity and recombinant protein yield are enhanced by
depletion of miR-7 activity via sponge decoy vectors", Biotechnology journal, 9(3),
pp. 396-404.
Sangokoya, C., Telen, M.J. & Chi, J.T. 2010, "microRNA miR-144 modulates oxidative
stress tolerance and associates with anemia severity in sickle cell disease", Blood,
vol. 116, no. 20, pp. 4338-4348.
Sauerwald, T.M., Betenbaugh, M.J. & Oyler, G.A. 2002a, "Inhibiting apoptosis in
mammalian cell culture using the caspase inhibitor XIAP and deletion mutants",
Biotechnology and bioengineering, vol. 77, no. 6, pp. 704-716.
Sauerwald, T.M., Betenbaugh, M.J. & Oyler, G.A. 2002b, "Inhibiting apoptosis in
mammalian cell culture using the caspase inhibitor XIAP and deletion mutants",
Biotechnology and bioengineering, vol. 77, no. 6, pp. 704-716.
Sauerwald, T.M., Figueroa, B.,Jr, Hardwick, J.M., Oyler, G.A. & Betenbaugh, M.J. 2006,
"Combining caspase and mitochondrial dysfunction inhibitors of apoptosis to limit
cell death in mammalian cell cultures", Biotechnology and bioengineering, vol. 94,
no. 2, pp. 362-372.
409
Sauerwald, T.M., Oyler, G.A. & Betenbaugh, M.J. 2003, "Study of caspase inhibitors for
limiting death in mammalian cell culture", Biotechnology and bioengineering, vol.
81, no. 3, pp. 329-340.
Saydam, O., Shen, Y., Wurdinger, T., Senol, O., Boke, E., James, M.F., Tannous, B.A.,
Stemmer-Rachamimov, A.O., Yi, M., Stephens, R.M., Fraefel, C., Gusella, J.F.,
Krichevsky, A.M. & Breakefield, X.O. 2009, "Downregulated microRNA-200a in
meningiomas promotes tumor growth by reducing E-cadherin and activating the
Wnt/beta-catenin signaling pathway", Molecular and cellular biology, vol. 29, no.
21, pp. 5923-5940.
Scheffer, A.R., Holdenrieder, S., Kristiansen, G., von Ruecker, A., Muller, S.C. &
Ellinger, J. 2012, "Circulating microRNAs in serum: novel biomarkers for patients
with bladder cancer?", World journal of urology, 32(2), pp. 353-8.
Schwarz, D.S., Hutvagner, G., Du, T., Xu, Z., Aronin, N. & Zamore, P.D. 2003,
"Asymmetry in the assembly of the RNAi enzyme complex", Cell, vol. 115, no. 2,
pp. 199-208.
Seitz, H. 2009, "Redefining microRNA targets", Current biology: CB, vol. 19, no. 10, pp.
870-873.
Sela-Culang, I., Kunik, V. & Ofran, Y. 2013, "The Structural Basis of Antibody-Antigen
Recognition", Frontiers in immunology, vol. 4, pp. 302.
Sengupta, N., Rose, S.T. & Morgan, J.A. 2011, "Metabolic flux analysis of CHO cell
metabolism in the late non-growth phase", Biotechnology and bioengineering, vol.
108, no. 1, pp. 82-92.
Serino, G., Sallustio, F., Cox, S.N., Pesce, F. & Schena, F.P. 2012, "Abnormal miR-148b
expression promotes aberrant glycosylation of IgA1 in IgA nephropathy", Journal
of the American Society of Nephrology: JASN, vol. 23, no. 5, pp. 814-824.
Sethupathy, P., Corda, B. & Hatzigeorgiou, A.G. 2006, "TarBase: A comprehensive
database of experimentally supported animal microRNA targets", RNA (New York,
N.Y.), vol. 12, no. 2, pp. 192-197.
Shahbazian, M.D. & Grunstein, M. 2007, "Functions of site-specific histone acetylation
and deacetylation", Annual Review of Biochemistry, vol. 76, pp. 75-100.
Shan, S.W., Fang, L., Shatseva, T., Rutnam, Z.J., Yang, X., Du, W., Lu, W.Y., Xuan,
J.W., Deng, Z. & Yang, B.B. 2013, "Mature miR-17-5p and passenger miR-17-3p
induce hepatocellular carcinoma by targeting PTEN, GalNT7 and vimentin in
different signal pathways", Journal of cell science, vol. 126, no. Pt 6, pp. 1517-1530.
Shen, D., Kiehl, T.R., Khattak, S.F., Li, Z.J., He, A., Kayne, P.S., Patel, V., Neuhaus,
I.M. & Sharfstein, S.T. 2010, "Transcriptomic responses to sodium chloride-induced
osmotic stress: a study of industrial fed-batch CHO cell cultures", Biotechnology
progress, vol. 26, no. 4, pp. 1104-1115.
410
Shen, Z., Zhan, G., Ye, D., Ren, Y., Cheng, L., Wu, Z. & Guo, J. 2012, "MicroRNA-34a
affects the occurrence of laryngeal squamous cell carcinoma by targeting the
antiapoptotic gene survivin", Medical oncology (Northwood, London, England), vol.
29, no. 4, pp. 2473-2480.
Shibanuma, M., Inoue, A., Ushida, K., Uchida, T., Ishikawa, F., Mori, K. & Nose, K.
2011, "Importance of mitochondrial dysfunction in oxidative stress response: A
comparative study of gene expression profiles", Free radical research, vol. 45, no.
6, pp. 672-680.
Shigenaga, M.K., Hagen, T.M. & Ames, B.N. 1994, "Oxidative damage and
mitochondrial decay in aging", Proceedings of the National Academy of Sciences of
the United States of America, vol. 91, no. 23, pp. 10771-10778.
Shin, A.H., Kil, I.S., Yang, E.S., Huh, T.L., Yang, C.H. & Park, J.W. 2004, "Regulation
of high glucose-induced apoptosis by mitochondrial NADP+-dependent isocitrate
dehydrogenase", Biochemical and biophysical research communications, vol. 325,
no. 1, pp. 32-38.
Shukla, G.C., Singh, J. & Barik, S. 2011, "MicroRNAs: Processing, Maturation, Target
Recognition and Regulatory Functions", Molecular and cellular pharmacology, vol.
3, no. 3, pp. 83-92.
Siberil, S., Dutertre, C.A., Fridman, W.H. & Teillaud, J.L. 2007, "FcgammaR: The key
to optimize therapeutic antibodies?", Critical reviews in oncology/hematology, vol.
62, no. 1, pp. 26-33.
Sibley, C.R., Seow, Y., Saayman, S., Dijkstra, K.K., El Andaloussi, S., Weinberg, M.S.
& Wood, M.J. 2012, "The biogenesis and characterization of mammalian
microRNAs of mirtron origin", Nucleic acids research, vol. 40, no. 1, pp. 438-448.
Siegel, C., Li, J., Liu, F., Benashski, S.E. & McCullough, L.D. 2011, "miR-23a regulation
of X-linked inhibitor of apoptosis (XIAP) contributes to sex differences in the
response to cerebral ischemia", Proceedings of the National Academy of Sciences of
the United States of America, vol. 108, no. 28, pp. 11662-11667.
Sikand, K., Slane, S.D. & Shukla, G.C. 2009, "Intrinsic expression of host genes and
intronic miRNAs in prostate carcinoma cells", Cancer cell international, vol. 9, pp.
21-2867-9-21.
Silber, J., Jacobsen, A., Ozawa, T., Harinath, G., Pedraza, A., Sander, C., Holland, E.C.
& Huse, J.T. 2012, "miR-34a repression in proneural malignant gliomas upregulates
expression of its target PDGFRA and promotes tumorigenesis", PloS one, vol. 7, no.
3, pp. e33844.
Silva, J.M., Li, M.Z., Chang, K., Ge, W., Golding, M.C., Rickles, R.J., Siolas, D., Hu, G.,
Paddison, P.J., Schlabach, M.R., Sheth, N., Bradshaw, J., Burchard, J., Kulkarni, A.,
Cavet, G., Sachidanandam, R., McCombie, W.R., Cleary, M.A., Elledge, S.J. &
Hannon, G.J. 2005, "Second-generation shRNA libraries covering the mouse and
human genomes", Nature genetics, vol. 37, no. 11, pp. 1281-1288.
411
Simmen, T., Lynes, E.M., Gesson, K. & Thomas, G. 2010, "Oxidative protein folding in
the endoplasmic reticulum: tight links to the mitochondria-associated membrane
(MAM)", Biochimica et biophysica acta, vol. 1798, no. 8, pp. 1465-1473.
Sinclair, A.M. & Elliott, S. 2005, "Glycoengineering: the effect of glycosylation on the
properties of therapeutic proteins", Journal of pharmaceutical sciences, vol. 94, no.
8, pp. 1626-1635.
Singh, S.K. 2011, "Impact of product-related factors on immunogenicity of
biotherapeutics", Journal of pharmaceutical sciences, vol. 100, no. 2, pp. 354-387.
Sleiman, R.J., Gray, P.P., McCall, M.N., Codamo, J. & Sunstrom, N.A. 2008,
"Accelerated cell line development using two-color fluorescence activated cell
sorting to select highly expressing antibody-producing clones", Biotechnology and
bioengineering, vol. 99, no. 3, pp. 578-587.
Smales, C.M., Dinnis, D.M., Stansfield, S.H., Alete, D., Sage, E.A., Birch, J.R., Racher,
A.J., Marshall, C.T. & James, D.C. 2004, "Comparative proteomic analysis of GS-
NS0 murine myeloma cell lines with varying recombinant monoclonal antibody
production rate", Biotechnology and bioengineering, vol. 88, no. 4, pp. 474-488.
Smith, A.L., Iwanaga, R., Drasin, D.J., Micalizzi, D.S., Vartuli, R.L., Tan, A.C. & Ford,
H.L. 2012, "The miR-106b-25 cluster targets Smad7, activates TGF-beta signaling,
and induces EMT and tumor initiating cell characteristics downstream of Six1 in
human breast cancer", Oncogene, 31(50), pp. 5162-71.
Song, E., Lee, S.K., Dykxhoorn, D.M., Novina, C., Zhang, D., Crawford, K., Cerny, J.,
Sharp, P.A., Lieberman, J., Manjunath, N. & Shankar, P. 2003, "Sustained small
interfering RNA-mediated human immunodeficiency virus type 1 inhibition in
primary macrophages", Journal of virology, vol. 77, no. 13, pp. 7174-7181.
Song, G., Zhang, Y. & Wang, L. 2009, "MicroRNA-206 targets notch3, activates
apoptosis, and inhibits tumor cell migration and focus formation", The Journal of
biological chemistry, vol. 284, no. 46, pp. 31921-31927.
Sood, P., Krek, A., Zavolan, M., Macino, G. & Rajewsky, N. 2006, "Cell-type-specific
signatures of microRNAs on target mRNA expression", Proceedings of the National
Academy of Sciences of the United States of America, vol. 103, no. 8, pp. 2746-2751.
Spanaki, C. & Plaitakis, A. 2012, "The role of glutamate dehydrogenase in mammalian
ammonia metabolism", Neurotoxicity research, vol. 21, no. 1, pp. 117-127.
Spenger, A., Ernst, W., Condreay, J.P., Kost, T.A. & Grabherr, R. 2004, "Influence of
promoter choice and trichostatin A treatment on expression of baculovirus delivered
genes in mammalian cells", Protein expression and purification, vol. 38, no. 1, pp.
17-23.
Steel, L.F. & Sanghvi, V.R. 2012, "Polycistronic expression of interfering RNAs from
RNA polymerase III promoters", Methods in molecular biology (Clifton, N.J.), vol.
815, pp. 347-359.
412
Stegmeier, F., Hu, G., Rickles, R.J., Hannon, G.J. & Elledge, S.J. 2005, "A lentiviral
microRNA-based system for single-copy polymerase II-regulated RNA interference
in mammalian cells", Proceedings of the National Academy of Sciences of the United
States of America, vol. 102, no. 37, pp. 13212-13217.
Stenvang, J., Petri, A., Lindow, M., Obad, S. & Kauppinen, S. 2012, "Inhibition of
microRNA function by antimiR oligonucleotides", Silence, vol. 3, no. 1, pp. 1.
Stiburek, L., Vesela, K., Hansikova, H., Pecina, P., Tesarova, M., Cerna, L., Houstek, J.
& Zeman, J. 2005, "Tissue-specific cytochrome c oxidase assembly defects due to
mutations in SCO2 and SURF1", The Biochemical journal, vol. 392, no. Pt 3, pp.
625-632.
Stoller, M.L., Chang, H.C. & Fekete, D.M. 2013, "Bicistronic gene transfer tools for
delivery of miRNAs and protein coding sequences", International journal of
molecular sciences, vol. 14, no. 9, pp. 18239-18255.
Strotbek, M., Florin, L., Koenitzer, J., Tolstrup, A., Kaufmann, H., Hausser, A. &
Olayioye, M.A. 2013, "Stable microRNA expression enhances therapeutic antibody
productivity of Chinese hamster ovary cells", Metabolic engineering, vol. 20, pp.
157-166.
Su, J., Zhang, A., Shi, Z., Ma, F., Pu, P., Wang, T., Zhang, J., Kang, C. & Zhang, Q. 2012,
"MicroRNA-200a suppresses the Wnt/beta-catenin signaling pathway by interacting
with beta-catenin", International journal of oncology, vol. 40, no. 4, pp. 1162-1170.
Sun, F., Fu, H., Liu, Q., Tie, Y., Zhu, J., Xing, R., Sun, Z. & Zheng, X. 2008a,
"Downregulation of CCND1 and CDK6 by miR-34a induces cell cycle arrest", FEBS
letters, vol. 582, no. 10, pp. 1564-1568.
Sun, F., Fu, H., Liu, Q., Tie, Y., Zhu, J., Xing, R., Sun, Z. & Zheng, X. 2008b,
"Downregulation of CCND1 and CDK6 by miR-34a induces cell cycle arrest", FEBS
letters, vol. 582, no. 10, pp. 1564-1568.
Sun, J., Zhou, M., Mao, Z. & Li, C. 2012, "Characterization and evolution of microRNA
genes derived from repetitive elements and duplication events in plants", PloS one,
vol. 7, no. 4, pp. e34092.
Sun, Q., Chen, X., Ma, J., Peng, H., Wang, F., Zha, X., Wang, Y., Jing, Y., Yang, H.,
Chen, R., Chang, L., Zhang, Y., Goto, J., Onda, H., Chen, T., Wang, M.R., Lu, Y.,
You, H., Kwiatkowski, D. & Zhang, H. 2011, "Mammalian target of rapamycin up-
regulation of pyruvate kinase isoenzyme type M2 is critical for aerobic glycolysis
and tumor growth", Proceedings of the National Academy of Sciences of the United
States of America, vol. 108, no. 10, pp. 4129-4134.
Sun, S., Li, W., Zhang, H., Zha, L., Xue, Y., Wu, X. & Zou, F. 2012a, "Requirement for
store-operated calcium entry in sodium butyrate-induced apoptosis in human colon
cancer cells", Bioscience reports, vol. 32, no. 1, pp. 83-90.
413
Sun, Y., Zhao, X., Zhou, Y. & Hu, Y. 2012b, "miR-124, miR-137 and miR-340 regulate
colorectal cancer growth via inhibition of the Warburg effect", Oncology reports,
vol. 28, no. 4, pp. 1346-1352.
Sung, Y.H. & Lee, G.M. 2005, "Enhanced human thrombopoietin production by sodium
butyrate addition to serum-free suspension culture of bcl-2-overexpressing CHO
cells", Biotechnology progress, vol. 21, no. 1, pp. 50-57.
Sung, Y.H., Lee, J.S., Park, S.H., Koo, J. & Lee, G.M. 2007, "Influence of co-down-
regulation of caspase-3 and caspase-7 by siRNAs on sodium butyrate-induced
apoptotic cell death of Chinese hamster ovary cells producing thrombopoietin",
Metabolic engineering, vol. 9, no. 5-6, pp. 452-464.
Sunley, K. & Butler, M. 2010, "Strategies for the enhancement of recombinant protein
production from mammalian cells by growth arrest", Biotechnology Advances, vol.
28, no. 3, pp. 385-394.
Sunley, K., Tharmalingam, T. & Butler, M. 2008a, "CHO cells adapted to hypothermic
growth produce high yields of recombinant beta-interferon", Biotechnology
progress, vol. 24, no. 4, pp. 898-906.
Sunley, K., Tharmalingam, T. & Butler, M. 2008b, "CHO cells adapted to hypothermic
growth produce high yields of recombinant beta-interferon", Biotechnology
progress, vol. 24, no. 4, pp. 898-906.
Suzuki, E., Niwa, R., Saji, S., Muta, M., Hirose, M., Iida, S., Shiotsu, Y., Satoh, M.,
Shitara, K., Kondo, M. & Toi, M. 2007a, "A nonfucosylated anti-HER2 antibody
augments antibody-dependent cellular cytotoxicity in breast cancer patients",
Clinical cancer research : an official journal of the American Association for Cancer
Research, vol. 13, no. 6, pp. 1875-1882.
Suzuki, E., Niwa, R., Saji, S., Muta, M., Hirose, M., Iida, S., Shiotsu, Y., Satoh, M.,
Shitara, K., Kondo, M. & Toi, M. 2007b, "A nonfucosylated anti-HER2 antibody
augments antibody-dependent cellular cytotoxicity in breast cancer patients",
Clinical cancer research : an official journal of the American Association for Cancer
Research, vol. 13, no. 6, pp. 1875-1882.
Sylvestre, Y., De Guire, V., Querido, E., Mukhopadhyay, U.K., Bourdeau, V., Major, F.,
Ferbeyre, G. & Chartrand, P. 2007, "An E2F/miR-20a autoregulatory feedback
loop", The Journal of biological chemistry, vol. 282, no. 4, pp. 2135-2143.
Tabuchi, H., Sugiyama, T., Tanaka, S. & Tainaka, S. 2010, "Overexpression of taurine
transporter in Chinese hamster ovary cells can enhance cell viability and product
yield, while promoting glutamine consumption", Biotechnology and bioengineering,
vol. 107, no. 6, pp. 998-1003.
Tagawa, H. & Seto, M. 2005, "A microRNA cluster as a target of genomic amplification
in malignant lymphoma", Leukemia, vol. 19, no. 11, pp. 2013-2016.
414
Tan, Y., Zhang, B., Wu, T., Skogerbo, G., Zhu, X., Guo, X., He, S. & Chen, R. 2009,
"Transcriptional inhibiton of Hoxd4 expression by miRNA-10a in human breast
cancer cells", BMC molecular biology, vol. 10, pp. 12-2199-10-12.
Tang, W., Zhu, J., Su, S., Wu, W., Liu, Q., Su, F. & Yu, F. 2012, "MiR-27 as a prognostic
marker for breast cancer progression and patient survival", PloS one, vol. 7, no. 12,
pp. e51702.
Tay, Y., Zhang, J., Thomson, A.M., Lim, B. & Rigoutsos, I. 2008, "MicroRNAs to
Nanog, Oct4 and Sox2 coding regions modulate embryonic stem cell
differentiation", Nature, vol. 455, no. 7216, pp. 1124-1128.
Templeton, N., Dean, J., Reddy, P. & Young, J.D. 2013, "Peak antibody production is
associated with increased oxidative metabolism in an industrially relevant fed-batch
CHO cell culture", Biotechnology and bioengineering, vol. 110, no. 7, pp. 2013-
2024.
Thaisuchat, H., Baumann, M., Pontiller, J., Hesse, F. & Ernst, W. 2011, "Identification
of a novel temperature sensitive promoter in CHO cells", BMC biotechnology, vol.
11, pp. 51-6750-11-51.
Tharmalingam, T., Sunley, K. & Butler, M. 2008, "High yields of monomeric
recombinant beta-interferon from macroporous microcarrier cultures under
hypothermic conditions", Biotechnology progress, vol. 24, no. 4, pp. 832-838.
Thomson, D.W., Bracken, C.P. & Goodall, G.J. 2011, "Experimental strategies for
microRNA target identification", Nucleic acids research, vol. 39, no. 16, pp. 6845-
6853.
Tigges, M. & Fussenegger, M. 2006, "Xbp1-based engineering of secretory capacity
enhances the productivity of Chinese hamster ovary cells", Metabolic engineering,
vol. 8, no. 3, pp. 264-272.
Tomasetti, M., Neuzil, J. & Dong, L. 2013, "MicroRNAs as regulators of mitochondrial
function: Role in cancer suppression", Biochimica et biophysica acta, 1840(4), pp.
1441-51.
Travers, K.J., Patil, C.K., Wodicka, L., Lockhart, D.J., Weissman, J.S. & Walter, P. 2000,
"Functional and genomic analyses reveal an essential coordination between the
unfolded protein response and ER-associated degradation", Cell, vol. 101, no. 3, pp.
249-258.
Trigona, W.L., Mullarky, I.K., Cao, Y. & Sordillo, L.M. 2006, "Thioredoxin reductase
regulates the induction of haem oxygenase-1 expression in aortic endothelial cells",
The Biochemical journal, vol. 394, no. Pt 1, pp. 207-216.
Truettner, J.S., Motti, D. & Dietrich, W.D. 2013, "MicroRNA overexpression increases
cortical neuronal vulnerability to injury", Brain research, vol. 1533, pp. 122-130.
415
Trummer, E., Fauland, K., Seidinger, S., Schriebl, K., Lattenmayer, C., Kunert, R.,
Vorauer-Uhl, K., Weik, R., Borth, N., Katinger, H. & Muller, D. 2006, "Process
parameter shifting: Part I. Effect of DOT, pH, and temperature on the performance
of Epo-Fc expressing CHO cells cultivated in controlled batch bioreactors",
Biotechnology and bioengineering, vol. 94, no. 6, pp. 1033-1044.
Tsao, Y.S., Cardoso, A.G., Condon, R.G., Voloch, M., Lio, P., Lagos, J.C., Kearns, B.G.
& Liu, Z. 2005, "Monitoring Chinese hamster ovary cell culture by the analysis of
glucose and lactate metabolism", Journal of Biotechnology, vol. 118, no. 3, pp. 316-
327.
Tseng, Y.C., Mozumdar, S. & Huang, L. 2009, "Lipid-based systemic delivery of
siRNA", Advanced Drug Delivery Reviews, vol. 61, no. 9, pp. 721-731.
Tsukamoto, Y., Nakada, C., Noguchi, T., Tanigawa, M., Nguyen, L.T., Uchida, T.,
Hijiya, N., Matsuura, K., Fujioka, T., Seto, M. & Moriyama, M. 2010, "MicroRNA-
375 is downregulated in gastric carcinomas and regulates cell survival by targeting
PDK1 and 14-3-3zeta", Cancer research, vol. 70, no. 6, pp. 2339-2349.
Ucakturk, E. 2012, "Analysis of glycoforms on the glycosylation site and the glycans in
monoclonal antibody biopharmaceuticals", Journal of separation science, vol. 35,
no. 3, pp. 341-350.
Uchiumi, T. & Kang, D. 2012, "The role of TFAM-associated proteins in mitochondrial
RNA metabolism", Biochimica et biophysica acta, vol. 1820, no. 5, pp. 565-570.
Urbich, C., Kaluza, D., Fromel, T., Knau, A., Bennewitz, K., Boon, R.A., Bonauer, A.,
Doebele, C., Boeckel, J.N., Hergenreider, E., Zeiher, A.M., Kroll, J., Fleming, I. &
Dimmeler, S. 2012, "MicroRNA-27a/b controls endothelial cell repulsion and
angiogenesis by targeting semaphorin 6A", Blood, vol. 119, no. 6, pp. 1607-1616.
Vaistij, F.E., Elias, L., George, G.L. & Jones, L. 2010, "Suppression of microRNA
accumulation via RNA interference in Arabidopsis thaliana", Plant Molecular
Biology, vol. 73, no. 4-5, pp. 391-397.
Vander Heiden, M.G., Cantley, L.C. & Thompson, C.B. 2009, "Understanding the
Warburg effect: the metabolic requirements of cell proliferation", Science (New
York, N.Y.), vol. 324, no. 5930, pp. 1029-1033.
Vella, M.C., Choi, E.Y., Lin, S.Y., Reinert, K. & Slack, F.J. 2004, "The C. elegans
microRNA let-7 binds to imperfect let-7 complementary sites from the lin-41
3'UTR", Genes & development, vol. 18, no. 2, pp. 132-137.
Ventura, A., Young, A.G., Winslow, M.M., Lintault, L., Meissner, A., Erkeland, S.J.,
Newman, J., Bronson, R.T., Crowley, D., Stone, J.R., Jaenisch, R., Sharp, P.A. &
Jacks, T. 2008, "Targeted deletion reveals essential and overlapping functions of the
miR-17 through 92 family of miRNA clusters", Cell, vol. 132, no. 5, pp. 875-886.
Vergoulis, T., Vlachos, I.S., Alexiou, P., Georgakilas, G., Maragkakis, M., Reczko, M.,
Gerangelos, S., Koziris, N., Dalamagas, T. & Hatzigeorgiou, A.G. 2012, "TarBase
416
6.0: capturing the exponential growth of miRNA targets with experimental support",
Nucleic acids research, vol. 40, no. Database issue, pp. D222-9.
Vogt, M., Munding, J., Gruner, M., Liffers, S.T., Verdoodt, B., Hauk, J., Steinstraesser,
L., Tannapfel, A. & Hermeking, H. 2011, "Frequent concomitant inactivation of
miR-34a and miR-34b/c by CpG methylation in colorectal, pancreatic, mammary,
ovarian, urothelial, and renal cell carcinomas and soft tissue sarcomas", Virchows
Archiv : an international journal of pathology, vol. 458, no. 3, pp. 313-322.
Walsh, G. 2010, "Biopharmaceutical benchmarks 2010", Nature biotechnology, vol. 28,
no. 9, pp. 917-924.
Walsh, G. & Jefferis, R. 2006, "Post-translational modifications in the context of
therapeutic proteins", Nature biotechnology, vol. 24, no. 10, pp. 1241-1252.
Wang, A., Zhang, B., Zhang, J., Wu, W. & Wu, W. 2013, "Embelin-induced brain glioma
cell apoptosis and cell cycle arrest via the mitochondrial pathway", Oncology
reports, vol. 29, no. 6, pp. 2473-2478.
Wang, A.H., Kruhlak, M.J., Wu, J., Bertos, N.R., Vezmar, M., Posner, B.I., Bazett-Jones,
D.P. & Yang, X.J. 2000, "Regulation of histone deacetylase 4 by binding of 14-3-3
proteins", Molecular and cellular biology, vol. 20, no. 18, pp. 6904-6912.
Wang, D.G., Sun, Y.B., Ye, F., Li, W., Kharbuja, P., Gao, L., Zhang, D.Y. & Suo, J.
2013, "Anti-tumor activity of the X-linked inhibitor of apoptosis (XIAP) inhibitor
embelin in gastric cancer cells", Molecular and cellular biochemistry, 386(1-2), pp.
143-52.
Wang, I.K., Hsieh, S.Y., Chang, K.M., Wang, Y.C., Chu, A., Shaw, S.Y., Ou, J.J. & Ho,
L. 2006, "A novel control scheme for inducing angiostatin-human IgG fusion protein
production using recombinant CHO cells in a oscillating bioreactor", Journal of
Biotechnology, vol. 121, no. 3, pp. 418-428.
Wang, R., Hu, Y., Song, G., Hao, C.J., Cui, Y., Xia, H.F. & Ma, X. 2012, "MiR-206
regulates neural cells proliferation and apoptosis via Otx2", Cellular physiology and
biochemistry : international journal of experimental cellular physiology,
biochemistry, and pharmacology, vol. 29, no. 3-4, pp. 381-390.
Wang, R., Wang, Z.X., Yang, J.S., Pan, X., De, W. & Chen, L.B. 2011a, "MicroRNA-
451 functions as a tumor suppressor in human non-small cell lung cancer by targeting
ras-related protein 14 (RAB14)", Oncogene, vol. 30, no. 23, pp. 2644-2658.
Wang, X., Xu, X., Ma, Z., Huo, Y., Xiao, Z., Li, Y. & Wang, Y. 2011b, "Dynamic
mechanisms for pre-miRNA binding and export by Exportin-5", RNA (New York,
N.Y.), vol. 17, no. 8, pp. 1511-1528.
Wang, X.F., Shi, Z.M., Wang, X.R., Cao, L., Wang, Y.Y., Zhang, J.X., Yin, Y., Luo, H.,
Kang, C.S., Liu, N., Jiang, T. & You, Y.P. 2012, "MiR-181d acts as a tumor
suppressor in glioma by targeting K-ras and Bcl-2", Journal of cancer research and
clinical oncology, vol. 138, no. 4, pp. 573-584.
417
Watson, W.H., Yang, X., Choi, Y.E., Jones, D.P. & Kehrer, J.P. 2004, "Thioredoxin and
its role in toxicology", Toxicological sciences : an official journal of the Society of
Toxicology, vol. 78, no. 1, pp. 3-14.
Weber, M., Baker, M.B., Moore, J.P. & Searles, C.D. 2010, "MiR-21 is induced in
endothelial cells by shear stress and modulates apoptosis and eNOS activity",
Biochemical and biophysical research communications, vol. 393, no. 4, pp. 643-648.
Wei, Y.Y., Naderi, S., Meshram, M., Budman, H., Scharer, J.M., Ingalls, B.P. &
McConkey, B.J. 2011, "Proteomics analysis of chinese hamster ovary cells
undergoing apoptosis during prolonged cultivation", Cytotechnology, vol. 63, no. 6,
pp. 663-677.
Welch, C., Chen, Y. & Stallings, R.L. 2007, "MicroRNA-34a functions as a potential
tumor suppressor by inducing apoptosis in neuroblastoma cells", Oncogene, vol. 26,
no. 34, pp. 5017-5022.
Wenham, R.M., Landt, M. & Easom, R.A. 1994, "Glucose activates the multifunctional
Ca2+/calmodulin-dependent protein kinase II in isolated rat pancreatic islets", The
Journal of biological chemistry, vol. 269, no. 7, pp. 4947-4952.
Winnard, P.,Jr, Glackin, C. & Raman, V. 2006, "Stable integration of an empty vector in
MCF-7 cells greatly alters the karyotype", Cancer genetics and cytogenetics, vol.
164, no. 2, pp. 174-176.
Wlaschin, K.F. & Hu, W.S. 2007, "Engineering cell metabolism for high-density cell
culture via manipulation of sugar transport", Journal of Biotechnology, vol. 131, no.
2, pp. 168-176.
Wong, D.C., Wong, K.T., Lee, Y.Y., Morin, P.N., Heng, C.K. & Yap, M.G. 2006,
"Transcriptional profiling of apoptotic pathways in batch and fed-batch CHO cell
cultures", Biotechnology and bioengineering, vol. 94, no. 2, pp. 373-382.
Wu, S.C. 2009a, "RNA interference technology to improve recombinant protein
production in Chinese hamster ovary cells", Biotechnology Advances, vol. 27, no. 4,
pp. 417-422.
Wu, S.C. 2009b, "RNA interference technology to improve recombinant protein
production in Chinese hamster ovary cells", Biotechnology Advances, vol. 27, no. 4,
pp. 417-422.
Wu, L., Fan, J. & Belasco, J.G. 2006, "MicroRNAs direct rapid deadenylation of mRNA",
Proceedings of the National Academy of Sciences of the United States of America,
vol. 103, no. 11, pp. 4034-4039.
Wu, S., Huang, S., Ding, J., Zhao, Y., Liang, L., Liu, T., Zhan, R. & He, X. 2010,
"Multiple microRNAs modulate p21Cip1/Waf1 expression by directly targeting its
3' untranslated region", Oncogene, vol. 29, no. 15, pp. 2302-2308.
418
Wu, S.C. 2009, "RNA interference technology to improve recombinant protein
production in Chinese hamster ovary cells", Biotechnology Advances, vol. 27, no. 4,
pp. 417-422.
Wurm, F.M. 2004, "Production of recombinant protein therapeutics in cultivated
mammalian cells", Nature biotechnology, vol. 22, no. 11, pp. 1393-1398.
Wurm, F.M. & Hacker, D. 2011, "First CHO genome", Nature biotechnology, vol. 29,
no. 8, pp. 718-720.
Xia, H.Q., Pan, Y., Peng, J. & Lu, G.X. 2011, "Over-expression of miR375 reduces
glucose-induced insulin secretion in Nit-1 cells", Molecular biology reports, vol. 38,
no. 5, pp. 3061-3065.
Xie, Y., Tobin, L.A., Camps, J., Wangsa, D., Yang, J., Rao, M., Witasp, E., Awad, K.S.,
Yoo, N., Ried, T. & Kwong, K.F. 2012, "MicroRNA-24 regulates XIAP to reduce
the apoptosis threshold in cancer cells", Oncogene, 9(32), pp. 2442-51.
Xu, C., Lu, Y., Pan, Z., Chu, W., Luo, X., Lin, H., Xiao, J., Shan, H., Wang, Z. & Yang,
B. 2007, "The muscle-specific microRNAs miR-1 and miR-133 produce opposing
effects on apoptosis by targeting HSP60, HSP70 and caspase-9 in cardiomyocytes",
Journal of cell science, vol. 120, no. Pt 17, pp. 3045-3052.
Xu, X., Nagarajan, H., Lewis, N.E., Pan, S., Cai, Z., Liu, X., Chen, W., Xie, M., Wang,
W., Hammond, S., Andersen, M.R., Neff, N., Passarelli, B., Koh, W., Fan, H.C.,
Wang, J., Gui, Y., Lee, K.H., Betenbaugh, M.J., Quake, S.R., Famili, I., Palsson,
B.O. & Wang, J. 2011, "The genomic sequence of the Chinese hamster ovary (CHO)-
K1 cell line", Nature biotechnology, vol. 29, no. 8, pp. 735-741.
Yamakuchi, M., Ferlito, M. & Lowenstein, C.J. 2008, "miR-34a repression of SIRT1
regulates apoptosis", Proceedings of the National Academy of Sciences of the United
States of America, vol. 105, no. 36, pp. 13421-13426.
Yamane-Ohnuki, N., Kinoshita, S., Inoue-Urakubo, M., Kusunoki, M., Iida, S., Nakano,
R., Wakitani, M., Niwa, R., Sakurada, M., Uchida, K., Shitara, K. & Satoh, M. 2004,
"Establishment of FUT8 knockout Chinese hamster ovary cells: an ideal host cell
line for producing completely defucosylated antibodies with enhanced antibody-
dependent cellular cytotoxicity", Biotechnology and bioengineering, vol. 87, no. 5,
pp. 614-622.
Yan, B., Zhao, L.H., Guo, J.T. & Zhao, J.L. 2012, "miR-429 regulation of osmotic stress
transcription factor 1 (OSTF1) in tilapia during osmotic stress", Biochemical and
biophysical research communications, vol. 426, no. 3, pp. 294-298.
Yan, D., Dong Xda, E., Chen, X., Wang, L., Lu, C., Wang, J., Qu, J. & Tu, L. 2009a,
"MicroRNA-1/206 targets c-Met and inhibits rhabdomyosarcoma development",
The Journal of biological chemistry, vol. 284, no. 43, pp. 29596-29604.
Yan, H.L., Xue, G., Mei, Q., Wang, Y.Z., Ding, F.X., Liu, M.F., Lu, M.H., Tang, Y., Yu,
H.Y. & Sun, S.H. 2009b, "Repression of the miR-17-92 cluster by p53 has an
419
important function in hypoxia-induced apoptosis", The EMBO journal, vol. 28, no.
18, pp. 2719-2732.
Yang, F., Li, Q.J., Gong, Z.B., Zhou, L., You, N., Wang, S., Li, X.L., Li, J.J., An, J.Z.,
Wang, D.S., He, Y. & Dou, K.F. 2014, "MicroRNA-34a targets Bcl-2 and sensitizes
human hepatocellular carcinoma cells to sorafenib treatment", Technology in cancer
research & treatment, vol. 13, no. 1, pp. 77-86.
Yang, H., Ye, D., Guan, K.L. & Xiong, Y. 2012, "IDH1 and IDH2 mutations in
tumorigenesis: mechanistic insights and clinical perspectives", Clinical cancer
research : an official journal of the American Association for Cancer Research, vol.
18, no. 20, pp. 5562-5571.
Yang, J.S., Phillips, M.D., Betel, D., Mu, P., Ventura, A., Siepel, A.C., Chen, K.C. & Lai,
E.C. 2011, "Widespread regulatory activity of vertebrate microRNA* species", RNA
(New York, N.Y.), vol. 17, no. 2, pp. 312-326.
Yang, M. & Butler, M. 2000, "Effects of ammonia on CHO cell growth, erythropoietin
production, and glycosylation", Biotechnology and bioengineering, vol. 68, no. 4,
pp. 370-380.
Yang, W.C., Lu, J., Nguyen, N.B., Zhang, A., Healy, N.V., Kshirsagar, R., Ryll, T. &
Huang, Y.M. 2014, "Addition of Valproic Acid to CHO Cell Fed-Batch Cultures
Improves Monoclonal Antibody Titers", Molecular biotechnology, vol. 56, no. 5, pp.
421-428.
Yang, X., Du, W.W., Li, H., Liu, F., Khorshidi, A., Rutnam, Z.J. & Yang, B.B. 2013,
"Both mature miR-17-5p and passenger strand miR-17-3p target TIMP3 and induce
prostate tumor growth and invasion", Nucleic acids research, 41(21), pp. 9688-704.
Yang, X., Rutnam, Z.J., Jiao, C., Wei, D., Xie, Y., Du, J., Zhong, L. & Yang, B.B. 2012,
"An anti-let-7 sponge decoys and decays endogenous let-7 functions", Cell cycle
(Georgetown, Tex.), vol. 11, no. 16, pp. 3097-3108.
Yee, J.C., de Leon Gatti, M., Philp, R.J., Yap, M. & Hu, W.S. 2008, "Genomic and
proteomic exploration of CHO and hybridoma cells under sodium butyrate
treatment", Biotechnology and bioengineering, vol. 99, no. 5, pp. 1186-1204.
Yee, J.C., Gerdtzen, Z.P. & Hu, W.S. 2009, "Comparative transcriptome analysis to
unveil genes affecting recombinant protein productivity in mammalian cells",
Biotechnology and bioengineering, vol. 102, no. 1, pp. 246-263.
Yeom, K.H., Lee, Y., Han, J., Suh, M.R. & Kim, V.N. 2006, "Characterization of
DGCR8/Pasha, the essential cofactor for Drosha in primary miRNA processing",
Nucleic acids research, vol. 34, no. 16, pp. 4622-4629.
Yoon, S.K., Song, J.Y. & Lee, G.M. 2003, "Effect of low culture temperature on specific
productivity, transcription level, and heterogeneity of erythropoietin in Chinese
hamster ovary cells", Biotechnology and bioengineering, vol. 82, no. 3, pp. 289-298.
420
Yoshikawa, T., Nakanishi, F., Ogura, Y., Oi, D., Omasa, T., Katakura, Y., Kishimoto, M.
& Suga, K.I. 2001, "Flow cytometry: an improved method for the selection of highly
productive gene-amplified CHO cells using flow cytometry", Biotechnology and
bioengineering, vol. 74, no. 5, pp. 435-442.
Young, J.D. 2013, "Metabolic flux rewiring in mammalian cell cultures", Current opinion
in biotechnology, vol. 24, no. 6, pp. 1108-1115.
Yuan, J.H., Yang, F., Chen, B.F., Lu, Z., Huo, X.S., Zhou, W.P., Wang, F. & Sun, S.H.
2011, "The histone deacetylase 4/SP1/microrna-200a regulatory network contributes
to aberrant histone acetylation in hepatocellular carcinoma", Hepatology (Baltimore,
Md.), vol. 54, no. 6, pp. 2025-2035.
Yun, C.Y., Liu, S., Lim, S.F., Wang, T., Chung, B.Y., Jiat Teo, J., Chuan, K.H., Soon,
A.S., Goh, K.S. & Song, Z. 2007, "Specific inhibition of caspase-8 and -9 in CHO
cells enhances cell viability in batch and fed-batch cultures", Metabolic engineering,
vol. 9, no. 5-6, pp. 406-418.
Zeng, Y. & Cullen, B.R. 2005, "Efficient processing of primary microRNA hairpins by
Drosha requires flanking nonstructured RNA sequences", The Journal of biological
chemistry, vol. 280, no. 30, pp. 27595-27603.
Zeng, Y. & Cullen, B.R. 2004, "Structural requirements for pre-microRNA binding and
nuclear export by Exportin 5", Nucleic acids research, vol. 32, no. 16, pp. 4776-
4785.
Zeng, Y. & Cullen, B.R. 2003, "Sequence requirements for micro RNA processing and
function in human cells", RNA (New York, N.Y.), vol. 9, no. 1, pp. 112-123.
Zhang, H., Hao, Y., Yang, J., Zhou, Y., Li, J., Yin, S., Sun, C., Ma, M., Huang, Y. & Xi,
J.J. 2011, "Genome-wide functional screening of miR-23b as a pleiotropic modulator
suppressing cancer metastasis", Nature communications, vol. 2, pp. 554.
Zhang, L., Liu, X., Jin, H., Guo, X., Xia, L., Chen, Z., Bai, M., Liu, J., Shang, X., Wu,
K., Pan, Y. & Fan, D. 2013, "miR-206 inhibits gastric cancer proliferation in part by
repressing cyclinD2", Cancer letters, vol. 332, no. 1, pp. 94-101.
Zhang, T. & Brown, J.H. 2004, "Role of Ca2+/calmodulin-dependent protein kinase II in
cardiac hypertrophy and heart failure", Cardiovascular research, vol. 63, no. 3, pp.
476-486.
Zhang, T., Kohlhaas, M., Backs, J., Mishra, S., Phillips, W., Dybkova, N., Chang, S.,
Ling, H., Bers, D.M., Maier, L.S., Olson, E.N. & Brown, J.H. 2007, "CaMKIIdelta
isoforms differentially affect calcium handling but similarly regulate HDAC/MEF2
transcriptional responses", The Journal of biological chemistry, vol. 282, no. 48, pp.
35078-35087.
Zhang, T., Liu, M., Wang, C., Lin, C., Sun, Y. & Jin, D. 2011, "Down-regulation of MiR-
206 promotes proliferation and invasion of laryngeal cancer by regulating VEGF
expression", Anticancer Research, vol. 31, no. 11, pp. 3859-3863.
421
Zhang, X., Hu, S., Zhang, X., Wang, L., Zhang, X., Yan, B., Zhao, J., Yang, A. & Zhang,
R. 2013, "MicroRNA-7 arrests cell cycle in G1 phase by directly targeting CCNE1
in human hepatocellular carcinoma cells", Biochemical and biophysical research
communications, 443(3), pp. 1078-84.
Zhao, S., Lin, Y., Xu, W., Jiang, W., Zha, Z., Wang, P., Yu, W., Li, Z., Gong, L., Peng,
Y., Ding, J., Lei, Q., Guan, K.L. & Xiong, Y. 2009, "Glioma-derived mutations in
IDH1 dominantly inhibit IDH1 catalytic activity and induce HIF-1alpha", Science
(New York, N.Y.), vol. 324, no. 5924, pp. 261-265.
Zheng, X., Baker, H., Hancock, W.S., Fawaz, F., McCaman, M. & Pungor, E.,Jr 2006,
"Proteomic analysis for the assessment of different lots of fetal bovine serum as a
raw material for cell culture. Part IV. Application of proteomics to the manufacture
of biological drugs", Biotechnology progress, vol. 22, no. 5, pp. 1294-1300.
Zhou, H., Liu, Z.G., Sun, Z.W., Huang, Y. & Yu, W.Y. 2010, "Generation of stable cell
lines by site-specific integration of transgenes into engineered Chinese hamster
ovary strains using an FLP-FRT system", Journal of Biotechnology, vol. 147, no. 2,
pp. 122-129.
Zhou, J., Li, Y.S., Nguyen, P., Wang, K.C., Weiss, A., Kuo, Y.C., Chiu, J.J., Shyy, J.Y.
& Chien, S. 2013, "Regulation of vascular smooth muscle cell turnover by
endothelial cell-secreted microRNA-126: role of shear stress", Circulation research,
vol. 113, no. 1, pp. 40-51.
Zhou, M., Crawford, Y., Ng, D., Tung, J., Pynn, A.F., Meier, A., Yuk, I.H.,
Vijayasankaran, N., Leach, K., Joly, J., Snedecor, B. & Shen, A. 2011a, "Decreasing
lactate level and increasing antibody production in Chinese Hamster Ovary cells
(CHO) by reducing the expression of lactate dehydrogenase and pyruvate
dehydrogenase kinases", Journal of Biotechnology, vol. 153, no. 1-2, pp. 27-34.
Zhou, M., Crawford, Y., Ng, D., Tung, J., Pynn, A.F., Meier, A., Yuk, I.H.,
Vijayasankaran, N., Leach, K., Joly, J., Snedecor, B. & Shen, A. 2011b, "Decreasing
lactate level and increasing antibody production in Chinese Hamster Ovary cells
(CHO) by reducing the expression of lactate dehydrogenase and pyruvate
dehydrogenase kinases", Journal of Biotechnology, vol. 153, no. 1-2, pp. 27-34.
Zhou, S.G., Wang, P., Pi, R.B., Gao, J., Fu, J.J., Fang, J., Qin, J., Zhang, H.J., Li, R.F.,
Chen, S.R., Tang, F.T. & Liu, P.Q. 2008, "Reduced expression of GSTM2 and
increased oxidative stress in spontaneously hypertensive rat", Molecular and cellular
biochemistry, vol. 309, no. 1-2, pp. 99-107.
Zhu, X., Santat, L.A., Chang, M.S., Liu, J., Zavzavadjian, J.R., Wall, E.A., Kivork, C.,
Simon, M.I. & Fraser, I.D. 2007, "A versatile approach to multiple gene RNA
interference using microRNA-based short hairpin RNAs", BMC molecular biology,
vol. 8, pp. 98.
422
Appendix
Figure A1: microRNAs used in primary screen
microRNA Pre-miR/Inhibitor (PM/AM)
miR-10a AM
miR-15a AM
miR-15b AM
miR-16-1 AM
miR-18a PM
miR-23b PM
miR-23b* PM
miR-25 AM
miR-27a AM
miR-29a AM
miR-34a PM
miR-34a AM
miR-34a* PM
miR-93 AM
miR-106b AM
miR-126 AM
miR-143 AM
miR-181d PM
mir-184 AM
miR-200a PM
miR-206 AM
miR-219 PM
miR-320 PM
miR-320a PM
miR-320a AM
miR-338-3p AM
miR-378 AM
miR-409-3p AM
miR-450 PM
miR-455-3p AM
miR-455-5p AM
miR-532-5p PM
miR-532-5p AM
miR-639 PM
miR-874 AM
AM = anti-miR for transient inhibition
PM = pre-miR (mimic) for transient over-expression
423
Figure A2: Relative data for transient transfection of anti-miR 126, -338-3p, -378 and -409-3p in CHO-K1-SEAP cells
Day 2 Day 4 Day 7 Day 2 Day 4 Day 7 Day 2 Day 4 Day 7
miR-126 1.03 1.06 0.8 0.95 1.1 0.88 0.92 1.04 1.09
1.22 0.86 0.86 1.02 0.97 0.79 0.83 1.11 0.92
1.06 1.4 1 0.93 1.03 1.07 0.87 0.73 0.99
0.6 0.47 1.33 0.76 0.82 1.1 1.25 1.43 0.82
0.57 0.49 1.07 0.69 0.74 1.19 1.19 1.48 1.11
1.18 1.13 1.59 0.96 0.9 1.03 0.81 0.8 0.64
0.88 0.9 0.66 0.96 0.92 0.87 1.09 1.02 1.3
0.77 0.59 0.83 0.72 0.86 1.21 0.92 1.45 1.12
0.99 1.05 0.99 0.91 0.95 1.15 0.92 0.9 1.16
0.9 0.8 1.18 0.92 0.83 0.93 1.02 1.03 0.78
1.29 1.97 1.12 0.83 1.16 0.93 0.64 0.59 0.83
0.87 0.87 1.22 0.89 1.04 1.11 1.02 1.19 0.91
1.19 0.87 0.83 1.22 1.36 1.44 1.01 1.55 1.71
1.08 1.27 0.95 1.09 1.09 1.08 1.01 0.85 1.13
0.99 1.26 1.5 0.92 0.96 1.09 0.89 0.76 0.73
0.86 0.98 0.6 0.94 0.97 0.95 1.09 1.09 1.57
0.92 0.9 0.65 0.98 1.16 0.84 1.06 1.28 1.29
miR-409-3p
Cell numbers/mL Protein Yield/mL Cell specific Productivity
miRNA
miR-338-3p
miR-378
424
Figure A3: High cell density transfection optimisation using siRNA against VCP
Figure A3: cells were transfected at a density of 4 x 105 cells/mL using 50 nM pre-miR-NC and siRNA
designed against VCP using a range of concentrations of INTERFERinTM of 1, 2 and 4 µl. Part A
shows cell numbers/mL while part B shows cell viability.
0.00E+00
2.00E+05
4.00E+05
6.00E+05
8.00E+05
1.00E+06
1.20E+06
1.40E+06
1.60E+06
1.80E+06
Cells PM-NC (1ul)
PM-NC (2ul)
PM-NC (4ul)
VCP (1ul)
VCP (2ul)
VCP (4ul)
Cells/mL
Day 2
80
82
84
86
88
90
92
94
96
98
100
Cells PM-NC (1ul)
PM-NC (2ul)
PM-NC (4ul)
VCP (1ul) VCP (2ul) VCP (4ul)
Viability %
Day 2
0.00E+00
2.00E+05
4.00E+05
6.00E+05
8.00E+05
1.00E+06
1.20E+06
1.40E+06
1.60E+06
1.80E+06
Cells PM-NC (1ul)
PM-NC (2ul)
PM-NC (4ul)
VCP (1ul)
VCP (2ul)
VCP (4ul)
Cells/mL
Day 2
80
82
84
86
88
90
92
94
96
98
100
Cells PM-NC (1ul)
PM-NC (2ul)
PM-NC (4ul)
VCP (1ul) VCP (2ul) VCP (4ul)
Viability %
Day 2
A
B
425
Figure A4: clustal alignment for miR-34 sponge sequence
Figure A4: Clustal alignment to confirm insertion of the miR-34 sponge sequence. Seq1 = template
sequence, Seq2 = sequence results from MWG Operon sequencing service. Star denotes a perfect
match.
Figure A5: single nucleotide difference between miR-23a and miR-23b
Figure A5: Sequence similarity between the mature miR-23a and miR-23b sequences. A single
nucleotide difference is indicated by a star (*) with the seed region of both miRNAs being highlighted
in a yellow box (nt 2-8).
miR-23b 5’-aucacauugccagggauuacc-3’
miR-23a 5’-aucacauugccagggauuucc-3’*
426
Figure A6: Diagnostic gels for miR-23b/-23b* sponge vector preparation
Figure A6: Diagnostic gels for miR-sponge preparation A) stock plasmid containing miR-23b/-23b*
(pEX-A) suitable for bacterial culture only B) Digestion of empty and negative control (NC) pd2-
EGFP-sponge with XhoI/EcoRI C) Extracted miR-23b/-23b* sequence inserts D) Mini-prep
diagnostic panel of miR-23b/-23b* sponge vectors E) Final diagnostic of Maxi-prep sponge vectors.
100bp
500bp
850bp
miR-23b* SpongemiR-23b Sponge
500 bp
650 bp
850 bp
100bp
200bp
100 bp
200 bp
300 bp
400 bp
500 bp
650 bp
850 bp1000 bp
1650 bp2000 bp
3000 bp
pEX-A
miR-23b
pEX-A
miR-23b*
A B C
D
E
427
Figure A7: Sequence alignment for miR-23b and miR-23b* sponge vectors
Figure A7: The Clustal sequence alignments for both miR-23b* (A) and miR-23b (B) sponge
sequences referenced against the original sequence designed. Seq1 denotes the original sequence
designed while Seq2 denotes the sequence information from MWG Operon. Perfect matches are
flagged with stars (*) while just the sponge sequence is highlighted in green.
Figure A8: pre-miR-23b hairpin with base-pair mismatches
Figure A8: This diagram was sourced from the miRBase database (http://www.mirbase.org/cgi-bin/mirna_entry.pl?acc=MI0000439) and demonstrates the structure of the pre-miR-23b hairpin upon
Drosha-mediated processing of the long pri-miRNA. We had specifically requested a tailor designed
miRNA (pre-miR-23b* mod) in which instead of an siRNA-like design of complete complementarity,
the base-pair mismatches are retained within its structure.
A miR-23b* sponge vector
B miR-23b sponge vector
miR-23b* (23b-5p)
miR-23b (23b-3p)
5’-
3’-
428
Figure A9: List of differentially expressed proteins when miR-23 and miR-NC
samples were randomised
Figure A9: When all clones were randomised and compared for differential protein expression, two
proteins were found to be DE. This was a quality control test as a means to determine the likelihood
of randomly identifying DE
Control Test FC
gi|344243859 2 120.07 0.02 1.86 Retinoic acid-induced protein 3 [Cricetulus griseus]
[MASS=39746]
8.74E+04 1.62E+05
1.853547
gi|344257799 1 53.63 0.03 5.66 Keratin, type II cytoskeletal 1b [Cricetulus griseus]
[MASS=93520]
1.22E+04 6.93E+04
5.680328
Description Average
Normalised
Abundances
Accessio
n
Peptides Score Anova
(p)*
Fold Tags
429
Figure A10: Differentially Expressed proteins from enriched mitochondrial (miR-23 sponge CHO-SEAP versus miR-NC sponge)
CS Anova (p) FC Gene I.D Description
65.21 9.61E-09 2.6 KIAA1881 Protein KIAA1881 [Cricetulus griseus] [MASS=39959]
473.97 2.52E-08 3.6 SBSN Suprabasin [Cricetulus griseus] [MASS=63932]
96.35 6.13E-07 3.7 GRN Granulins [Cricetulus griseus] [MASS=63924]
158.9 1.41E-06 1.7 LIMA1 LIM domain and actin-binding protein 1 [Cricetulus griseus] [MASS=84545]
76.55 1.52E-06 6.2 AHNAK2 Protein AHNAK2 [Cricetulus griseus] [MASS=81741]
204.15 2.19E-06 2.6 EPRS Bifunctional aminoacyl-tRNA synthetase [Cricetulus griseus] [MASS=154896]
290.49 2.30E-06 1.9 CALD1 Non-muscle caldesmon [Cricetulus griseus] [MASS=61921]
77.33 2.38E-06 1.5 HTATSF1 HIV Tat-specific factor 1-like [Cricetulus griseus] [MASS=60645]
62.16 2.49E-06 2.5 HSPG Basement membrane-specific heparan sulfate proteoglycan core protein [Cricetulus griseus] [MASS=334137]
47.16 3.08E-06 2.2 NIP7 60S ribosome subunit biogenesis protein NIP7-like [Cricetulus griseus] [MASS=10402]
42.96 3.50E-06 2.9 ADA Adenosine deaminase [Cricetulus griseus] [MASS=40940]
110.25 5.57E-06 1.5 HMGB2 High mobility group protein B2 [Cricetulus griseus] [MASS=18197]
216.74 6.04E-06 1.7 NOP2 Putative ribosomal RNA methyltransferase NOP2 [Cricetulus griseus] [MASS=86249]
42.91 6.49E-06 2.2 MCM3 DNA replication licensing factor MCM3 [Cricetulus griseus] [MASS=90790]
83.37 7.49E-06 2.3 DLGAP5 Disks large-associated protein 5 [Cricetulus griseus] [MASS=63127]
60.3 9.07E-06 2.8 ZCCHC3 Zinc finger CCHC domain-containing protein 3 [Cricetulus griseus] [MASS=30279]
112.46 9.29E-06 6.9 AKAP12 A-kinase anchor protein 12 [Cricetulus griseus] [MASS=91046]
499.19 9.69E-06 2.1 RCN1 Reticulocalbin-3 [Cricetulus griseus] [MASS=34399]
172.09 1.06E-05 1.6 SCP2 Non-specific lipid-transfer protein [Cricetulus griseus] [MASS=58892]
528.27 1.09E-05 2.4 GSTM6 Glutathione S-transferase Mu 6 [Cricetulus griseus] [MASS=86559]
85.95 1.19E-05 2.4 KIF2C Kinesin-like protein KIF2C
107.61 1.97E-05 1.7 P4HA1 Prolyl 4-hydroxylase subunit alpha-1 [Cricetulus griseus] [MASS=58084]
293.89 2.13E-05 2 CALU Calumenin [Cricetulus griseus] [MASS=87917]
84.54 2.25E-05 1.8 CAPRIN1 Caprin-1 [Cricetulus griseus] [MASS=73689]
40.54 2.28E-05 2.9 PTER Phosphotriesterase-related protein [Cricetulus griseus] [MASS=39278]
430
284.33 2.41E-05 2.5 GAA Lysosomal alpha-glucosidase [Cricetulus griseus] [MASS=105847]
56.39 2.55E-05 1.8 UAP1L1 UDP-N-acetylhexosamine pyrophosphorylase-like protein 1 [Cricetulus griseus] [MASS=58175]
44.26 2.57E-05 23.2 FABP4 Fatty acid-binding protein, adipocyte [Cricetulus griseus] [MASS=14753]
415.32 2.66E-05 1.6 ITGB1 Integrin beta-1 [Cricetulus griseus] [MASS=88342]
96.32 2.70E-05 1.7 CTSB Cathepsin B [Cricetulus griseus] [MASS=37504]
40.09 2.92E-05 1.8 TXNRD1 Thioredoxin reductase 1, cytoplasmic [Cricetulus griseus] [MASS=61858]
50.25 3.29E-05 2.7 MDH1 Malate dehydrogenase, cytoplasmic [Cricetulus griseus] [MASS=36466]
419.27 3.33E-05 1.6 GP50 GP50 [Cricetulus griseus] [MASS=46598]
152.86 3.40E-05 1.6 IQGAP1 Ras GTPase-activating-like protein IQGAP1 [Cricetulus griseus] [MASS=191377]
160.75 3.46E-05 1.6 SMARCA5 SWI/SNF-related matrix-associated actin-dependent regulator of chromatin subfamily A member 5
57.31 3.49E-05 2.5 DPP2 Dipeptidyl-peptidase 2 [Cricetulus griseus] [MASS=56397]
99.72 4.63E-05 2 IPO5 Importin-5 [Cricetulus griseus] [MASS=89713]
85.75 4.73E-05 2 CLINT1 Clathrin interactor 1 [Cricetulus griseus] [MASS=68215]
576.94 4.78E-05 1.5 HSP90B1 Endoplasmin [Cricetulus griseus] [MASS=92622]
85.86 5.00E-05 1.9 SH3BGRL3 SH3 domain-binding glutamic acid-rich-like protein [Cricetulus griseus] [MASS=12747]
134.41 5.03E-05 1.6 RRBP1 Ribosome-binding protein 1 [Cricetulus griseus] [MASS=216696]
42.91 5.12E-05 2.1 SKIV2L2 Superkiller viralicidic activity 2-like 2 [Cricetulus griseus] [MASS=96202]
43.27 5.35E-05 1.4 EIF3H Eukaryotic translation initiation factor 3 subunit H [Cricetulus griseus] [MASS=34239]
173.71 5.48E-05 1.6 LPH Lactase-phlorizin hydrolase [Cricetulus griseus] [MASS=301280]
155.61 5.81E-05 1.7 SHMT Serine hydroxymethyltransferase, mitochondrial [Cricetulus griseus] [MASS=55908]
420.16 5.99E-05 1.7 TOP2A RecName: Full=DNA topoisomerase 2-alpha; AltName: Full=DNA topoisomerase II, alpha isozyme
48.89 6.04E-05 3.6 GUSB Beta-glucuronidase [Cricetulus griseus] [MASS=74804]
60.27 6.20E-05 1.9 PRPF8 Pre-mRNA-processing-splicing factor 8 [Cricetulus griseus] [MASS=30142]
287.95 6.33E-05 1.4 STIP1 Stress-induced-phosphoprotein 1
461.22 6.40E-05 1.4 ANXA1 Annexin A1 [Cricetulus griseus] [MASS=38861]
991.99 6.58E-05 1.8 CALR Calreticulin [Cricetulus griseus] [MASS=48241]
155.37 6.71E-05 1.9 VCP Transitional endoplasmic reticulum ATPase [Cricetulus griseus] [MASS=237642]
780.39 6.72E-05 1.4 GAPDH Glyceraldehyde-3-phosphate dehydrogenase
48.83 7.11E-05 3.1 ALCAM CD166 antigen [Cricetulus griseus] [MASS=65083]
431
42.71 7.35E-05 1.5 VCL Vinculin [Cricetulus griseus] [MASS=124877]
534.81 7.65E-05 1.8 PDIA4 Protein disulfide-isomerase A4 [Cricetulus griseus] [MASS=72586]
43.2 8.00E-05 2.5 TWSG1 Twisted gastrulation protein-like 1 [Cricetulus griseus] [MASS=24747]
177.24 8.07E-05 1.7 HP1 heterochromatin protein 1 alpha [Cricetulus griseus]
485.82 8.17E-05 1.5 MYH10 Myosin-10 [Cricetulus griseus] [MASS=229037]
77.41 8.38E-05 1.7 CNN2 Calponin-2 [Cricetulus griseus] [MASS=32507]
42.27 8.41E-05 1.4 EHD3 EH domain-containing protein 3 [Cricetulus griseus] [MASS=60915]
69.02 8.73E-05 6 RPL7 60S ribosomal protein L7 [Cricetulus griseus] [MASS=29237]
998.19 9.15E-05 1.6 NUMA1 Nuclear mitotic apparatus protein 1 [Cricetulus griseus] [MASS=340229]
67.81 0.000104 1.6 Proline synthetase co-transcribed bacterial-like protein [Cricetulus griseus] [MASS=24790]
130.72 0.000106 2 PFDN6 prefoldin subunit 6 [Mus musculus]
106.24 0.000108 1.7 TXNDC5 Thioredoxin domain-containing protein 5 [Cricetulus griseus] [MASS=46354]
86.63 0.00011 1.8 PLAA Phospholipase A-2-activating protein [Cricetulus griseus] [MASS=87183]
74.08 0.000117 1.5 NCAPH Condensin complex subunit 2 [Cricetulus griseus] [MASS=81308]
652.03 0.000121 1.6 PDIA3 ERP57 protein [Cricetulus griseus] [MASS=56796]
47.4 0.000122 1.4 GRSF1 G-rich sequence factor 1 [Cricetulus griseus] [MASS=70369]
83.65 0.000128 3 H2AFV histone H2A.V isoform 1 [Homo sapiens]
40.47 0.000135 1.6 HLAA MHC class I antigen Hm1-C1 [Cricetulus griseus] [MASS=38752]
154.98 0.000136 2 HMOX1 Heme oxygenase 1 [Cricetulus griseus] [MASS=33023]
313.51 0.00014 2.8 FASN Fatty acid synthase [Cricetulus griseus] [MASS=271857]
42.21 0.000144 1.9 ZNF593 Zinc finger protein 593 [Cricetulus griseus] [MASS=15247]
113.62 0.00018 2.1 MCM5 DNA replication licensing factor MCM5 [Cricetulus griseus] [MASS=82347]
261.37 0.000194 1.4 CBX3 chromobox protein homolog 3 [Homo sapiens]
94.68 0.000198 1.3 AK2 Adenylate kinase 2, mitochondrial [Cricetulus griseus] [MASS=15424]
135.15 0.000205 2.2 MFGE8 Lactadherin [Cricetulus griseus] [MASS=16152]
40.63 0.000211 2.3 GNGT2 guanine nucleotide-binding protein G(I)/G(S)/G(O) subunit gamma-2 precursor [Mus musculus]
80.18 0.000223 1.8 FUS RNA-binding protein FUS [Cricetulus griseus] [MASS=21204]
53.68 0.000229 1.2 RPS27L ribosomal protein S27-like, isoform CRA_a [Homo sapiens]
119.96 0.000234 1.4 SRSF3 Splicing factor, arginine/serine-rich 3 [Cricetulus griseus] [MASS=19310]
432
60.94 0.000236 3.1 KIF4 Chromosome-associated kinesin KIF4 [Cricetulus griseus] [MASS=185097]
524.15 0.000241 1.6 PDIA3 Protein disulfide-isomerase A3 [Cricetulus griseus] [MASS=50051]
632.16 0.000241 1.7 MYBBP1A Myb-binding protein 1A [Cricetulus griseus] [MASS=105149]
53.85 0.000255 2 SLC7A1 High affinity cationic amino acid transporter 1 [Cricetulus griseus] [MASS=67614]
62.52 0.000264 2.9 HSPB1 Heat shock protein beta-1 [Cricetulus griseus] [MASS=8903]
106.62 0.000265 1.6 RRP15 RRP15-like protein [Cricetulus griseus] [MASS=42329]
1075.25 0.000265 1.4 GRP78 78 kDa glucose-regulated protein binding
756.55 0.000268 1.3 HSP60 60 kDa heat shock protein, mitochondrial
57.17 0.000269 1.5 COPB1 Coatomer subunit beta' [Cricetulus griseus] [MASS=89506]
105.9 0.000275 1.7 HDAC2 Histone deacetylase 2 [Cricetulus griseus] [MASS=51941]
107.47 0.000316 1.8 MCM2 DNA replication licensing factor MCM2 [Cricetulus griseus] [MASS=102184]
107.76 0.000326 1.5 VAPB Vesicle-associated membrane protein-associated protein B [Cricetulus griseus] [MASS=27502]
128.89 0.000329 1.4 DNTTIP2 Deoxynucleotidyltransferase terminal-interacting protein 2 [Cricetulus griseus] [MASS=85134]
136.67 0.000342 1.4 NUFIP2 Nuclear fragile X mental retardation-interacting protein 2 [Cricetulus griseus] [MASS=56644]
67.07 0.00035 1.5 NUP93 Nuclear pore complex protein Nup93 [Cricetulus griseus] [MASS=89109]
47.74 0.00035 1.7 KIF15 Kinesin-like protein KIF15 [Cricetulus griseus] [MASS=125621]
123.84 0.000367 1.9 CAD CAD protein [Cricetulus griseus] [MASS=243029]
145.15 0.000398 1.3 CTTN Src substrate cortactin [Cricetulus griseus] [MASS=61469]
109.1 0.000407 1.5 OXCT1 Succinyl-CoA:3-ketoacid-coenzyme A transferase 1, mitochondrial [Cricetulus griseus] [MASS=56218]
46.59 0.000411 3.5 PSAP Sulfated glycoprotein 1 [Cricetulus griseus] [MASS=27374]
319.39 0.000414 1.9 SND1 nuclease domain-containing protein 1 [Cricetulus griseus] [MASS=99748]
96.49 0.000417 1.8 WDR43 WD repeat-containing protein 43 [Cricetulus griseus] [MASS=55890]
180.99 0.000423 1.6 ATP1A1 Sodium/potassium-transporting ATPase subunit alpha-1 [Cricetulus griseus] [MASS=84781]
120.34 0.000436 1.4 MAP4 Microtubule-associated protein 4 [Cricetulus griseus] [MASS=138221]
221.57 0.000439 1.4 NSFL1C NSFL1 cofactor p47 [Cricetulus griseus] [MASS=41009]
121 0.000454 2.1 PLS3 RecName: Full=Plastin-3; AltName: Full=T-plastin
87.48 0.00046 1.6 CYF1P1 Cytoplasmic FMR1-interacting protein 1 [Cricetulus griseus] [MASS=145275]
71.91 0.00046 1.7 ATP2A3 Sarcoplasmic/endoplasmic reticulum calcium ATPase 3 [Cricetulus griseus] [MASS=103607]
52.46 0.000473 1.8 UHRF1 E3 ubiquitin-protein ligase UHRF1 [Cricetulus griseus] [MASS=88095]
433
45.14 0.000476 1.6 GMPS GMP synthase [glutamine-hydrolyzing] [Cricetulus griseus] [MASS=130016]
109.81 0.000483 1.3 NPM1 Nucleophosmin [Cricetulus griseus] [MASS=31052]
97.59 0.000487 1.5 MTHFD2 Bifunctional methylenetetrahydrofolate dehydrogenase/cyclohydrolase, mitochondrial [Cricetulus griseus]
69.28 0.000491 1.4 DDX6 PREDICTED: probable ATP-dependent RNA helicase DDX6 isoform 1 [Sus scrofa]
110.09 0.000492 1.4 PSMD2 26S proteasome non-ATPase regulatory subunit 2 [Cricetulus griseus] [MASS=72983]
197.29 0.000505 1.4 ANXA2 Annexin A2 [Cricetulus griseus] [MASS=27037]
65.4 0.000535 2.3 HSPH1 RecName: Full=Heat shock protein 105 kDa; AltName: Full=Heat shock 110 kDa protein
493.84 0.000553 1.4 SDHA Succinate dehydrogenase [ubiquinone] flavoprotein subunit, mitochondrial [Cricetulus griseus] [MASS=71658]
56.16 0.000556 1.3 RRM1 Ribonucleoside-diphosphate reductase large subunit [Cricetulus griseus] [MASS=90217]
324.58 0.000565 2.2 MYH11 Myosin-11 [Cricetulus griseus] [MASS=227190]
71.8 0.000572 2.2 GSTA3 Glutathione S-transferase alpha-3 [Cricetulus griseus] [MASS=14752]
93.72 0.000575 2.4 ABCF1 ATP-binding cassette sub-family F member 1 [Cricetulus griseus] [MASS=65859]
121.17 0.000581 1.5 PSME1 Proteasome activator complex subunit 1 [Cricetulus griseus] [MASS=24514]
434.44 0.000603 1.7 DDX1 RNA helicase [Mesocricetus auratus]
269.45 0.000617 1.7 CLTC Clathrin heavy chain 1 [Cricetulus griseus] [MASS=223961]
80.08 0.000637 1.6 SLC7A8 Large neutral amino acids transporter small subunit 1 [Cricetulus griseus] [MASS=55205]
168.98 0.00067 1.6 HYOU1 Hypoxia up-regulated protein 1
274.94 0.000693 1.6 LRPPRC Leucine-rich PPR motif-containing protein, mitochondrial [Cricetulus griseus] [MASS=158425]
40.28 0.000698 1.6 HNRNPUL1 Heterogeneous nuclear ribonucleoprotein U-like protein 2 [Cricetulus griseus] [MASS=72420]
294.31 0.000726 1.5 FLNB Filamin-B [Cricetulus griseus] [MASS=277844]
60.42 0.000737 2.3 BCL9L B-cell CLL/lymphoma 9-like protein [Cricetulus griseus] [MASS=156723]
77.91 0.000745 1.4 RHOA transforming protein RhoA precursor [Rattus norvegicus]
103.35 0.000751 1.5 RAB7A Ras-related protein Rab-7a [Cricetulus griseus] [MASS=23518]
51.97 0.000763 1.9 CETN3 Centrin-3 [Cricetulus griseus] [MASS=19558]
58.48 0.000795 1.3 CCT7 T-complex protein 1 subunit eta [Cricetulus griseus] [MASS=54906]
77.13 0.000817 1.4 PFDN4 Prefoldin subunit 4 [Cricetulus griseus] [MASS=15226]
276.02 0.000827 1.3 KIF27 Kinesin-like protein KIF27 [Cricetulus griseus] [MASS=185707]
183.23 0.000829 1.8 SLC25A3 Phosphate carrier protein, mitochondrial [Cricetulus griseus] [MASS=26018]
113.17 0.000839 2.6 SPRR1A Cornifin-A [Cricetulus griseus] [MASS=12457]
434
50.66 0.000873 1.6 EIF1AY eukaryotic translation initiation factor 1A [Mus musculus]
121.03 0.000876 1.3 LYAR Cell growth-regulating nucleolar protein [Cricetulus griseus] [MASS=43169]
55.92 0.000896 1.9 PPA1 Inorganic pyrophosphatase [Cricetulus griseus] [MASS=36997]
40.96 0.0009 1.9 CAND1 Cullin-associated NEDD8-dissociated protein 1 [Cricetulus griseus] [MASS=133622]
173.13 0.000923 1.4 RPL6 60S ribosomal protein L6 [Cricetulus griseus] [MASS=32777]
47.96 0.000938 1.9 L7RN6 lethal, Chr 7, Rinchik 6 [Mus musculus]
175.92 0.000989 1.4 YWHAZ 14-3-3 protein zeta/delta [Cricetulus griseus] [MASS=27715]
336.44 0.001001 1.3 ALDOA Fructose-bisphosphate aldolase A [Cricetulus griseus] [MASS=39382]
160.65 0.001043 1.4 TES Testin [Cricetulus griseus] [MASS=47209]
249.77 0.001078 1.4 EIF4G1 Eukaryotic translation initiation factor 4 gamma 1 [Cricetulus griseus] [MASS=175081]
211.82 0.001097 1.3 TMPO Lamina-associated polypeptide 2, isoforms alpha/zeta [Cricetulus griseus] [MASS=34881]
141.92 0.001111 1.6 KPNB1 Importin subunit beta-1 [Cricetulus griseus] [MASS=211197]
42.51 0.001122 1.7 YWHAH 14-3-3 protein eta [Cricetulus griseus] [MASS=27084]
41.01 0.001125 1.7 HDDC2 HD domain-containing protein 3 [Cricetulus griseus] [MASS=20271]
49.9 0.001205 1.8 SLC3A2 putative CD98 protein [Cricetulus griseus] [MASS=58926]
200.38 0.001206 1.3 LDHA LDH-A [Cricetulus griseus] [MASS=36519]
234.43 0.001211 1.4 HMGCS1 Hydroxymethylglutaryl-CoA synthase, cytoplasmic
48.07 0.001228 1.7 HSPA4 Heat shock 70 kDa protein 4 [Cricetulus griseus] [MASS=62990]
753.03 0.001268 1.4 EEF2 elongation factor 2 [Cricetulus griseus] [MASS=95323]
306.33 0.001271 1.2 TPI Triosephosphate isomerase [Cricetulus griseus] [MASS=20407]
371.2 0.001338 1.3 HSPA9 Stress-70 protein, mitochondrial
74.66 0.001374 1.5 HGPRT RecName: Full=Hypoxanthine-guanine phosphoribosyltransferase; Short=HGPRT; Short=HGPRTase
369.15 0.001428 1.4 GSTP1 RecName: Full=Glutathione S-transferase P; AltName: Full=GST class-pi
114.51 0.001441 1.7 MTHFD1 C-1-tetrahydrofolate synthase, cytoplasmic [Cricetulus griseus] [MASS=101286]
146.47 0.001474 1.4 RCN1 Reticulocalbin-1 [Cricetulus griseus] [MASS=29316]
56.71 0.001485 1.4 ANLN Actin-binding protein anillin [Cricetulus griseus] [MASS=122923]
230.11 0.001584 1.7 NOLC1 Nucleolar phosphoprotein p130 [Cricetulus griseus] [MASS=73455]
61.69 0.001609 1.6 TCOF1 Treacle protein [Cricetulus griseus] [MASS=127813]
263.85 0.001643 1.4 NUDC Nuclear migration protein nudC [Cricetulus griseus] [MASS=32771]
435
160.48 0.001645 1.5 MAP1B Microtubule-associated protein 1B [Cricetulus griseus] [MASS=260385]
152.73 0.001718 1.3 WBP2 WW domain-binding protein 2 [Cricetulus griseus] [MASS=27992]
639.78 0.001771 1.4 PDI Prolyl 4-hydroxylase subunit beta
117.81 0.00179 1.3 IDH3A Isocitrate dehydrogenase [NAD] subunit alpha, mitochondrial [Cricetulus griseus] [MASS=39684]
87.63 0.001806 1.5 LGALS1 Galectin-1
188.09 0.001807 1.4 PRKCSH Glucosidase 2 subunit beta [Cricetulus griseus] [MASS=62353]
177.85 0.001833 1.3 BST2 Bone marrow stromal antigen 2
70.34 0.001869 1.3 TRIM28 Transcription intermediary factor 1-beta [Cricetulus griseus] [MASS=58046]
63.76 0.001889 1.3 DDX10 putative ATP-dependent RNA helicase DDX10 [Cricetulus griseus] [MASS=101337]
69.21 0.001909 2 AIM2 Interferon-inducible protein [Cricetulus griseus] [MASS=14990]
315.2 0.001918 1.2 ACTB Actin, cytoplasmic 1
84.3 0.001949 1.3 PPP2R1A Serine/threonine-protein phosphatase 2A 65 kDa regulatory subunit A beta isoform [Cricetulus griseus]
63.6 0.001956 1.6 MATR3 Matrin-3 [Cricetulus griseus] [MASS=94625]
53.37 0.001985 1.3 SMARCC1 SWI/SNF complex subunit SMARCC1 [Cricetulus griseus] [MASS=154765]
143.8 0.001998 1.2 HNRNPAB Heterogeneous nuclear ribonucleoprotein A/B [Cricetulus griseus] [MASS=23754]
240.68 0.002132 1.5 ACO2 Aconitate hydratase, mitochondrial [Cricetulus griseus] [MASS=85466]
85.16 0.002167 1.5 SNX1 Sorting nexin-1 [Cricetulus griseus] [MASS=58922]
56.33 0.002223 1.8 LEPRE1 Prolyl 3-hydroxylase 1 [Cricetulus griseus] [MASS=73130]
89.12 0.002264 2.3 HMGN1 Nucleosome-binding protein 1 [Cricetulus griseus] [MASS=33339]
85 0.002286 1.4 MRPS7 28S ribosomal protein S7, mitochondrial [Cricetulus griseus] [MASS=17257]
53.95 0.002298 1.4 SUCLA2 Succinyl-CoA ligase [ADP-forming] subunit beta, mitochondrial [Cricetulus griseus] [MASS=25797]
59.63 0.002313 1.4 MAGOH protein mago nashi homolog [Homo sapiens]
94.58 0.002324 1.4 DDX56 putative ATP-dependent RNA helicase DDX56 [Cricetulus griseus] [MASS=61201]
41.04 0.002333 1.5 HAT1 Histone acetyltransferase type B catalytic subunit [Cricetulus griseus] [MASS=21390]
41.44 0.002371 1.4 ERP29 Endoplasmic reticulum protein ERp29 [Cricetulus griseus] [MASS=28603]
215.81 0.002435 1.3 NME2 Nucleoside diphosphate kinase B [Cricetulus griseus] [MASS=17340]
71.77 0.002584 1.5 GB2 growth factor receptor-bound protein 2 isoform 1 [Homo sapiens]
46.64 0.002633 1.6 NUP205 Nuclear pore complex protein Nup205 [Cricetulus griseus] [MASS=223157]
53.37 0.002636 1.4 TALDO1 RecName: Full=Transaldolase
436
189.83 0.002673 1.3 RPS15 40S ribosomal protein S15 [Cricetulus griseus] [MASS=12990]
164.35 0.00269 1.5 TOP1 DNA topoisomerase I [Cricetulus griseus] [MASS=90837]
146.99 0.002708 1.5 HDGF Hepatoma-derived growth factor [Cricetulus griseus] [MASS=22716]
48.73 0.002736 1.5 SUMF2 Sulfatase-modifying factor 2 [Cricetulus griseus] [MASS=7750]
46.99 0.002781 1.5 SRSF7 Splicing factor, arginine/serine-rich 7 [Cricetulus griseus] [MASS=26153]
89.77 0.002846 2.3 CMAS N-acylneuraminate cytidylyltransferase [Cricetulus griseus] [MASS=31274]
96.89 0.002873 6.6 SLC25A5 ADP/ATP translocase 2 [Cricetulus griseus] [MASS=46576]
206.18 0.002913 1.3 RPSA 40S ribosomal protein SA
86.59 0.002921 1.5 RANGAP1 Ran GTPase-activating protein 1 [Cricetulus griseus] [MASS=63142]
263.02 0.003013 1.5 ARHGDIA Rho GDP-dissociation inhibitor 1 [Cricetulus griseus] [MASS=23423]
231.19 0.003016 1.3 HSD17B10 3-hydroxyacyl-CoA dehydrogenase type-2 [Cricetulus griseus] [MASS=27174]
203.53 0.003086 2.3 RPN1 Dolichyl-diphosphooligosaccharide--protein glycosyltransferase subunit 1 [Cricetulus griseus] [MASS=68577]
832.3 0.003147 1.4 MYH9 Myosin-9 [Cricetulus griseus] [MASS=225794]
203.2 0.003181 1.2 TPR Nucleoprotein TPR [Cricetulus griseus] [MASS=238753]
57.3 0.003233 1.3 ARD1 hypothetical protein I79_013960 [Cricetulus griseus] [MASS=25829]
108.67 0.00327 1.4 GATAD2B Transcriptional repressor p66-beta [Cricetulus griseus] [MASS=64918]
136.53 0.003279 2 SF3B3 splicing factor 3B subunit 3 [Mus musculus]
111.74 0.003378 1.3 RPS18 40S ribosomal protein S18 [Cricetulus griseus] [MASS=17763]
58.08 0.003381 1.6 KHDRBS1 KH domain-containing, RNA-binding, signal transduction-associated protein 1 [Cricetulus griseus] [MASS=48326]
75.88 0.003612 1.4 UBXN1 UBX domain-containing protein 1 [Cricetulus griseus] [MASS=23179]
75.04 0.003631 1.7 MFGE8 Lactadherin [Cricetulus griseus] [MASS=22114]
52.21 0.003653 1.6 SDF2L1 Stromal cell-derived factor 2-like protein 1 [Cricetulus griseus] [MASS=19011]
45.27 0.003722 1.8 COX5A Cytochrome c oxidase subunit 5A, mitochondrial [Cricetulus griseus] [MASS=8603]
171.03 0.003763 1.5 SPTAN1 Spectrin alpha chain, brain [Cricetulus griseus] [MASS=258831]
271.84 0.003789 1.3 SFPQ splicing factor proline/glutamine rich (polypyrimidine tract binding protein associated) [Rattus norvegicus]
196.86 0.003789 1.4 RBM14 RNA-binding protein 14 [Cricetulus griseus] [MASS=50151]
48.78 0.003814 1.5 LAMC1 Laminin subunit gamma-1 [Cricetulus griseus] [MASS=172121]
47.43 0.003845 1.5 SLC1A5 Neutral amino acid transporter B(0) [Cricetulus griseus] [MASS=17378]
78.34 0.003922 1.3 ADPRH [Protein ADP-ribosylarginine] hydrolase [Cricetulus griseus] [MASS=38789]
437
195.07 0.003951 1.5 LRPAP1 Alpha-2-macroglobulin receptor-associated protein [Cricetulus griseus] [MASS=41173]
280.78 0.004029 1.2 HDLBP Vigilin [Cricetulus griseus] [MASS=141720]
43.14 0.004073 1.9 Cab45 45 kDa calcium-binding protein [Cricetulus griseus] [MASS=41237]
140.74 0.00413 1.9 RPL3 60S ribosomal protein L3 [Cricetulus griseus] [MASS=46152]
87.27 0.004148 1.3 SF3B1 splicing factor 3B subunit 1 isoform 1 [Homo sapiens]
304.02 0.004163 1.3 DDX17 DEAD (Asp-Glu-Ala-Asp) box polypeptide 17, isoform CRA_a [Rattus norvegicus]
140.66 0.004164 1.3 GANAB Neutral alpha-glucosidase AB [Cricetulus griseus] [MASS=106977]
40.99 0.004195 2.9 LAMP1 Lysosome-associated membrane glycoprotein 1
162.94 0.004226 1.3 BAG3 BAG family molecular chaperone regulator 3 [Cricetulus griseus] [MASS=61111]
101.43 0.00423 1.4 ATX10 Ataxin-10 [Cricetulus griseus] [MASS=34265]
131.59 0.004301 1.3 MDH2 Malate dehydrogenase, mitochondrial [Cricetulus griseus] [MASS=28696]
91.91 0.004348 1.6 ABCG3 ATP-binding cassette sub-family G member 3 [Cricetulus griseus] [MASS=69556]
139.2 0.004449 1.3 HNRNPM Heterogeneous nuclear ribonucleoprotein M [Cricetulus griseus] [MASS=60304]
80.15 0.004493 1.4 EWSR1 RNA-binding protein EWS [Cricetulus griseus] [MASS=68536]
41.01 0.004497 1.3 RAB1B ras-related protein Rab-1B [Mus musculus]
65.48 0.00451 1.6 SET Protein SET [Cricetulus griseus] [MASS=44274]
195.18 0.004615 1.5 FKBP4 FK506-binding protein 4 [Cricetulus griseus] [MASS=46625]
44.57 0.004662 2 IDI1 Isopentenyl-diphosphate Delta-isomerase 1 [Cricetulus griseus] [MASS=32694]
123.1 0.004826 1.5 UQCRC1 Cytochrome b-c1 complex subunit 1, mitochondrial [Cricetulus griseus] [MASS=40615]
199.47 0.004925 1.4 RPL7A 60S ribosomal protein L7a [Cricetulus griseus] [MASS=22240]
43.78 0.005069 1.4 DNAJC8 DnaJ-like subfamily C member 8 [Cricetulus griseus] [MASS=29840]
185.23 0.005078 1.6 EEF1B2 Elongation factor 1-beta [Cricetulus griseus] [MASS=24687]
142.75 0.005097 1.8 PSMD12 26S proteasome non-ATPase regulatory subunit 12 [Cricetulus griseus] [MASS=52978]
56.01 0.005112 1.5 DDB1 DNA damage-binding protein 1 [Cricetulus griseus] [MASS=126926]
150.21 0.005145 1.3 RPS13 RecName: Full=40S ribosomal protein S13
379.43 0.005165 1.5 MKI67 Antigen KI-67 [Cricetulus griseus] [MASS=356150]
59.13 0.005188 1.6 PGAM1 Phosphoglycerate mutase 1 [Cricetulus griseus] [MASS=20205]
57.83 0.005192 1.3 WNK1 Serine/threonine-protein kinase WNK1 [Cricetulus griseus] [MASS=157436]
124.49 0.005193 1.5 GFM1 Elongation factor G, mitochondrial [Cricetulus griseus] [MASS=83298]
438
1139.28 0.005193 1.5 LMNA Lamin-A/C [Cricetulus griseus] [MASS=64041]
57.24 0.005195 1.5 CEBPZ CCAAT/enhancer-binding protein zeta [Cricetulus griseus] [MASS=120496]
112.2 0.005225 1.4 KGD1 2-oxoglutarate dehydrogenase E1 component, mitochondrial [Cricetulus griseus] [MASS=104514]
64.63 0.005231 1.4 SYAP1 Synapse-associated protein 1 [Cricetulus griseus] [MASS=35619]
394.01 0.005336 1.8 PDIA6 Protein disulfide-isomerase A6 [Cricetulus griseus] [MASS=28401]
121.4 0.005428 1.3 CACYBP Calcyclin-binding protein [Cricetulus griseus] [MASS=21260]
57.72 0.005504 8.6 HIST2H2AA3 histone H2A type 2-A [Homo sapiens]
41.28 0.005535 1.3 RPS16 40S ribosomal protein S16 [Cricetulus griseus] [MASS=15699]
120.44 0.005659 1.3 EIF3G Eukaryotic translation initiation factor 3 subunit G [Cricetulus griseus] [MASS=35716]
72.63 0.005743 1.2 AIFM1 Apoptosis-inducing factor 1, mitochondrial [Cricetulus griseus] [MASS=66780]
40.65 0.005753 2 SPTBN1 Spectrin beta chain, brain 1 [Cricetulus griseus] [MASS=94523]
251.05 0.005771 1.4 UBA1 Ubiquitin-like modifier-activating enzyme 1 [Cricetulus griseus] [MASS=117697]
95.07 0.005797 1.8 HNRNPU Heterogeneous nuclear ribonucleoprotein U [Cricetulus griseus] [MASS=57620]
102.32 0.005826 1.8 DDX21 Nucleolar RNA helicase 2 [Cricetulus griseus] [MASS=80626]
94.1 0.005877 1.4 RPL7 60S ribosomal protein L7 [Cricetulus griseus] [MASS=17054]
51.38 0.005887 1.5 RPL29 60S ribosomal protein L29 [Cricetulus griseus] [MASS=12365]
43.74 0.005956 1.4 PMPCB Mitochondrial-processing peptidase subunit beta [Cricetulus griseus] [MASS=40055]
42.85 0.005957 1.3 ANXA5 Annexin A5 [Cricetulus griseus] [MASS=36065]
398.31 0.005973 1.4 LAML1 Lamin-L(I) [Cricetulus griseus] [MASS=55093]
76.63 0.006009 1.4 TMED9 Transmembrane emp24 domain-containing protein 9 [Cricetulus griseus] [MASS=26765]
48.69 0.006101 13.9 TPM2 Tropomyosin beta chain [Cricetulus griseus] [MASS=30686]
91.36 0.006151 1.3 EBP2 putative rRNA-processing protein EBP2 [Cricetulus griseus] [MASS=35026]
189.08 0.006175 1.3 ATAD3A ATPase family AAA domain-containing protein 3 [Cricetulus griseus] [MASS=66291]
73.25 0.006177 2.1 SGTA Small glutamine-rich tetratricopeptide repeat-containing protein alpha [Cricetulus griseus] [MASS=34090]
56.98 0.006188 1.7 EIF2S3 Eukaryotic translation initiation factor 2 subunit 3 [Cricetulus griseus] [MASS=51107]
44.36 0.006279 1.4 DDX39A ATP-dependent RNA helicase DDX39 [Cricetulus griseus] [MASS=48709]
68.25 0.006284 1.3 RPS2 40S ribosomal protein S2 [Cricetulus griseus] [MASS=8599]
88.54 0.006377 1.5 PSMD6 26S proteasome non-ATPase regulatory subunit 6 [Cricetulus griseus] [MASS=45566]
55.02 0.006636 3.1 HNRNPC Heterogeneous nuclear ribonucleoproteins C1/C2 [Cricetulus griseus] [MASS=24495]
439
55.1 0.006649 1.5 UBXN4 UBX domain-containing protein 4 [Cricetulus griseus] [MASS=56719]
107.31 0.006721 1.2 ICAM1 intracellular adhesion molecule 1 [Cricetulus griseus]
73.61 0.006835 1.3 UBE2V1 ubiquitin-conjugating enzyme E2 K isoform 1 [Homo sapiens]
72.75 0.006847 1.3 HEXIM1 Protein HEXIM1 [Cricetulus griseus] [MASS=40208]
81.13 0.006848 1.4 PSAT1 Phosphoserine aminotransferase [Cricetulus griseus] [MASS=33926]
195.33 0.007068 3.5 PABPC1 Polyadenylate-binding protein 1 [Cricetulus griseus] [MASS=62714]
249.85 0.007155 3 HSP90AA1 Heat shock protein HSP 90-alpha [Cricetulus griseus] [MASS=38095]
164.82 0.007238 1.4 NCL nucleolin, C23 [Cricetulus griseus] [MASS=73289]
86.37 0.007316 1.5 G6PD RecName: Full=Glucose-6-phosphate 1-dehydrogenase; Short=G6PD
67.1 0.007333 1.2 6PGD 6-phosphogluconate dehydrogenase, decarboxylating [Cricetulus griseus] [MASS=53376]
160.83 0.007423 1.4 TCP1 T-complex protein 1 subunit alpha
135.5 0.007429 1.2 RPL9 60S ribosomal protein L9 [Cricetulus griseus] [MASS=22652]
73.09 0.007567 1.5 MTAP S-methyl-5'-thioadenosine phosphorylase [Cricetulus griseus] [MASS=35587]
344.42 0.00761 1.4 HSP90AB1 Heat shock protein HSP 90-beta [Cricetulus griseus] [MASS=47807]
48.81 0.007636 1.5 PUR3 phosphoribosylglycinamide transformylase [Cricetulus griseus] [MASS=107446]
136.73 0.007637 1.2 SF3B1 Splicing factor 3 subunit 1 [Cricetulus griseus] [MASS=88710]
79.02 0.007671 1.1 RPL19 60S ribosomal protein L19 [Cricetulus griseus] [MASS=6891]
144.98 0.007927 1.3 SRRM2 Serine/arginine repetitive matrix protein 2 [Cricetulus griseus] [MASS=304044]
40.1 0.00794 7.6 RPL7A 60S ribosomal protein L7a [Cricetulus griseus] [MASS=15698]
46.96 0.008026 1.4 CPSF6 Cleavage and polyadenylation specificity factor subunit 6 [Cricetulus griseus] [MASS=53622]
90.99 0.008104 1.6 ARCN1 coatomer subunit delta [Mus musculus]
110.87 0.008124 1.4 ACTN4 Alpha-actinin-4 [Cricetulus griseus] [MASS=26219]
105.25 0.008335 1.8 HNRNPH2 heterogeneous nuclear ribonucleoprotein H2 [Mus musculus]
61.77 0.008415 1.3 CCT3 T-complex protein 1 subunit gamma [Cricetulus griseus] [MASS=60620]
44.31 0.008612 2.2 DYNC1H1 Cytoplasmic dynein 1 heavy chain 1 [Cricetulus griseus] [MASS=526782]
60.58 0.008695 3 PSMD14 26S proteasome non-ATPase regulatory subunit 14 [Homo sapiens]
145.99 0.008978 1.5 RPS17 40S ribosomal protein S17 [Mus musculus]
137.39 0.009137 1.4 ILF3 Interleukin enhancer-binding factor 3 [Cricetulus griseus] [MASS=97550]
109.8 0.009137 1.4 VARS Valyl-tRNA synthetase [Cricetulus griseus] [MASS=166356]
440
42.45 0.009183 1.5 DTYMK Thymidylate kinase [Cricetulus griseus] [MASS=8404]
54.54 0.009237 1.4 CCT4 T-complex protein 1 subunit delta [Cricetulus griseus] [MASS=42136]
87.01 0.009261 1.9 CRK Proto-oncogene C-crk [Cricetulus griseus] [MASS=36171]
87.63 0.009353 1.4 RPS21 40S ribosomal protein S21 [Cricetulus griseus] [MASS=9143]
49.08 0.009364 1.4 PLEKHO2 Pleckstrin-likey domain-containing family O member 2 [Cricetulus griseus] [MASS=54045]
468.75 0.009537 1.4 HSP90AA1 RecName: Full=Heat shock protein HSP 90-alpha
45.68 0.009552 1.7 MTA1 Metastasis-associated protein MTA1 [Cricetulus griseus] [MASS=79077]
45.42 0.009559 1.4 RAB14 ras-related protein Rab-14 [Mus musculus]
112.55 0.009801 1.4 PPIB Peptidyl-prolyl cis-trans isomerase B [Cricetulus griseus] [MASS=23634]
50.37 0.010036 1.4 NACA Nascent polypeptide-associated complex subunit alpha, muscle-specific form [Cricetulus griseus] [MASS=34787]
47.59 0.010052 1.7 RPL10 60S ribosomal protein L10 [Cricetulus griseus] [MASS=23428]
103 0.010084 1.6 RPL5 60S ribosomal protein L5 [Cricetulus griseus] [MASS=34413]
117.94 0.010379 1.5 HNRNPA1 Putative heterogeneous nuclear ribonucleoprotein A1-like protein 3 [Cricetulus griseus] [MASS=35945]
231.34 0.010501 1.3 MSN Moesin [Cricetulus griseus] [MASS=34529]
49.53 0.011035 2.2 BCLAF1 Bcl-2-associated transcription factor 1 [Cricetulus griseus] [MASS=47008]
66.36 0.011279 1.2 RPS18 40S ribosomal protein S18 [Cricetulus griseus] [MASS=18872]
72.87 0.011304 1.3 LGALS3 Galectin-3
105.82 0.011314 1.3 RPS26 40S ribosomal protein S26 [Rattus norvegicus]
56.93 0.011323 1.3 KIF5B Kinesin-1 heavy chain [Cricetulus griseus] [MASS=87850]
62.94 0.011644 1.5 SMC1A Structural maintenance of chromosomes protein 1A [Cricetulus griseus] [MASS=143220]
79.67 0.011724 2.1 GPRC5A Retinoic acid-induced protein 3 [Cricetulus griseus] [MASS=39746]
44.13 0.011798 1.3 CTSL Cathepsin L1 [Cricetulus griseus] [MASS=37248]
54.27 0.011871 1.1 PSMD4 26S proteasome non-ATPase regulatory subunit 4 [Cricetulus griseus] [MASS=41091]
43.31 0.011876 2.8 CAT Catalase [Cricetulus griseus] [MASS=63099]
53.31 0.01223 1.3 GSR Glutathione reductase, mitochondrial [Cricetulus griseus] [MASS=44949]
46.24 0.012371 1.5 ACTN4 Alpha-actinin-4 [Cricetulus griseus] [MASS=37673]
243.53 0.012419 1.5 NME1 Nucleoside diphosphate kinase A [Cricetulus griseus] [MASS=17195]
77.74 0.012505 1.4 CHMP4B Charged multivesicular body protein 4b [Cricetulus griseus] [MASS=21459]
62.52 0.01279 1.4 AGFG1 Arf-GAP domain and FG repeats-containing protein 1 [Cricetulus griseus] [MASS=50397]
441
250.63 0.012804 1.4 DDX5 probable ATP-dependent RNA helicase DDX5 [Rattus norvegicus]
60.6 0.012926 1.5 ZC3HAV1 Zinc finger CCCH-type antiviral protein 1 [Cricetulus griseus] [MASS=110826]
81.25 0.013009 1.9 EFTUD2 116 kDa U5 small nuclear ribonucleoprotein component [Cricetulus griseus] [MASS=109491]
101.18 0.013245 1.3 PES1 Pescadillo-like [Cricetulus griseus] [MASS=41566]
57.81 0.013468 1.3 RPS14 40S ribosomal protein S14 [Homo sapiens]
47.92 0.013665 2.4 NPTN Neuroplastin [Cricetulus griseus] [MASS=25086]
51.4 0.013744 1.4 CTNND1 Catenin delta-1 [Cricetulus griseus] [MASS=104051]
522.43 0.014198 2.1 HIST1H2AG Histone H2A type 1 [Cricetulus griseus] [MASS=34185]
45.77 0.014308 3 TKT Transketolase [Cricetulus griseus] [MASS=67716]
57.48 0.0145 1.6 LMNB2 Lamin-B2 [Cricetulus griseus] [MASS=67733]
62.86 0.014539 1.5 ADSL adenylosuccinate lyase [Cricetulus griseus] [MASS=54955]
89.69 0.014654 1.6 CCT8 T-complex protein 1 subunit theta [Cricetulus griseus] [MASS=22223]
75.9 0.014659 1.6 LBR Lamin-B receptor [Cricetulus griseus] [MASS=70428]
40.42 0.014831 1.3 MCM4 DNA replication licensing factor MCM4 [Cricetulus griseus] [MASS=96542]
100.56 0.015125 1.4 AKR1A1 Alcohol dehydrogenase [NADP+] [Cricetulus griseus] [MASS=36510]
41.99 0.01528 1.4 BOLA2 BolA-like protein 1 [Cricetulus griseus] [MASS=13002]
200.78 0.015326 1.8 SPTBN1 Spectrin beta chain, brain 1 [Cricetulus griseus] [MASS=166268]
40.24 0.015595 2.2 PA2G4 Proliferation-associated protein 2G4 [Cricetulus griseus] [MASS=43715]
43.83 0.015626 1.8 NPC2 Epididymal secretory protein E1 [Cricetulus griseus] [MASS=16471]
54.24 0.015636 1.3 RPL30 60S ribosomal protein L30 [Cricetulus griseus] [MASS=14716]
100.35 0.015764 1.3 STOML2 Stomatin-like protein 2 [Cricetulus griseus] [MASS=38199]
669.08 0.015962 1.4 FLNA Filamin-A [Cricetulus griseus] [MASS=263654]
65.53 0.016153 1.4 SNAP29 Synaptosomal-associated protein 29 [Cricetulus griseus] [MASS=21734]
171.7 0.016198 1.4 YBX3 DNA-binding protein A [Cricetulus griseus] [MASS=11426]
61.81 0.016287 1.4 UCHL5 Ubiquitin carboxyl-terminal hydrolase isozyme L5 [Cricetulus griseus] [MASS=37660]
65.39 0.016512 1.2 ACTR2 Actin-related protein 2 [Cricetulus griseus] [MASS=20222]
65.72 0.016568 1.4 GPC1 Glypican-1 [Cricetulus griseus] [MASS=86871]
242.47 0.016589 1.4 HNRNPA3 Heterogeneous nuclear ribonucleoprotein A3 [Cricetulus griseus] [MASS=26953]
58.36 0.016636 1.3 EIF3C Eukaryotic translation initiation factor 3 subunit C [Cricetulus griseus] [MASS=23144]
442
45.22 0.016895 1.4 PSMD11 26S proteasome non-ATPase regulatory subunit 11 [Cricetulus griseus] [MASS=29618]
44.45 0.01736 1.2 RPA2 Replication protein A 32 kDa subunit [Cricetulus griseus] [MASS=27524]
63.14 0.017467 3.2 TARDBP TAR DNA-binding protein 43 [Cricetulus griseus] [MASS=22576]
76.47 0.017845 1.4 FXR1 RecName: Full=Fragile X mental retardation syndrome-related protein 1
65.37 0.017889 1.3 BAIAP2 Brain-specific angiogenesis inhibitor 1-associated protein 2
58.73 0.018158 1.5 TBL2 Transducin beta-like protein 2 [Cricetulus griseus] [MASS=39587]
455.92 0.018196 1.2 HNRNPA2B1 Heterogeneous nuclear ribonucleoproteins A2/B1 [Cricetulus griseus] [MASS=29047]
53.28 0.018361 1.1 LMAN2 Vesicular integral-membrane protein VIP36 [Cricetulus griseus] [MASS=40227]
178.72 0.018574 1.9 GSTM5 Glutathione S-transferase Mu 5 [Cricetulus griseus] [MASS=26967]
355.42 0.018585 1.4 HIST1H2AG Histone H2A type 1 [Cricetulus griseus] [MASS=28463]
69.98 0.018808 1.5 PDLIM1 PDZ and LIM domain protein 1 [Cricetulus griseus] [MASS=35971]
40.29 0.018817 1.3 OPA1 Dynamin-like 120 kDa protein, mitochondrial [Cricetulus griseus] [MASS=117444]
59.2 0.019116 1.5 RRAS2 Ras-related protein R-Ras2 [Cricetulus griseus] [MASS=18078]
42.24 0.019157 1.9 UBA52 PREDICTED: ubiquitin B-like [Rattus norvegicus]
85.81 0.019179 1.2 RPL23A 60S ribosomal protein L23a [Cricetulus griseus] [MASS=17655]
114.63 0.019394 1.5 TUBA1C tubulin alpha-1C chain [Mus musculus]
426.28 0.019401 1.2 PLEC Plectin
110.18 0.019683 1.3 PEX19 Peroxisomal biogenesis factor 19
115.43 0.019901 3.4 HNRNPD Heterogeneous nuclear ribonucleoprotein D0 [Cricetulus griseus] [MASS=29283]
57.01 0.019938 2.4 S100A13 Protein S100-A13 [Cricetulus griseus] [MASS=11241]
46.34 0.020284 1.6 SUCLG1 Succinyl-CoA ligase [GDP-forming] subunit alpha, mitochondrial [Cricetulus griseus] [MASS=34895]
45.85 0.020412 1.2 ERG11 Lanosterol 14-alpha demethylase [Cricetulus griseus] [MASS=56767]
64.37 0.020911 1.2 GPS1 COP9 signalosome complex subunit 1 [Cricetulus griseus] [MASS=53947]
74.72 0.021424 1.4 RNH1 Ribonuclease inhibitor [Cricetulus griseus] [MASS=50022]
77.01 0.021542 1.2 SES1 Seryl-tRNA synthetase, cytoplasmic
60.49 0.021554 1.7 RPL27A 60S ribosomal protein L27a [Bos taurus]
132.1 0.021709 1.4 RPS3A 40S ribosomal protein S3a [Rattus norvegicus]
86.93 0.021721 1.3 SF3B4 splicing factor 3B subunit 4 [Mus musculus]
92.41 0.021732 1.6 SLC25A5 ADP/ATP translocase 2 [Cricetulus griseus] [MASS=29247]
443
50.91 0.022257 1.4 RBM28 RNA-binding protein 28 [Cricetulus griseus] [MASS=87436]
73.45 0.022335 1.2 TPI Triosephosphate isomerase [Cricetulus griseus] [MASS=16434]
74.38 0.022832 1.3 CSPG4 Chondroitin sulfate proteoglycan 4 [Cricetulus griseus] [MASS=252010]
135.26 0.022923 1.7 RPL4 60S ribosomal protein L4 [Cricetulus griseus] [MASS=32179]
74.08 0.022936 1.7 NAT10 N-acetyltransferase 10 [Cricetulus griseus] [MASS=115604]
44.37 0.023737 1.4 FKBP10 FK506-binding protein 10 [Cricetulus griseus] [MASS=64718]
58.11 0.023846 1.2 SEPT7 Septin-7 [Cricetulus griseus] [MASS=41218]
56.84 0.024636 1.5 ATP2B3 Plasma membrane calcium-transporting ATPase 3 [Cricetulus griseus] [MASS=130571]
41.35 0.024777 1.7 S100A6 protein S100-A6 [Mus musculus]
99.45 0.024964 1.6 KHSRP Far upstream element-binding protein 2 [Cricetulus griseus] [MASS=62198]
154.83 0.025369 1.2 YWHAQ 14-3-3 protein theta [Cricetulus griseus] [MASS=27749]
45.84 0.026109 1.6 RCN2 Reticulocalbin-2 [Cricetulus griseus] [MASS=37132]
42.38 0.026185 1.3 SAFB Scaffold attachment factor B1 [Cricetulus griseus] [MASS=102833]
45.15 0.026507 1.9 HNRNPG Heterogeneous nuclear ribonucleoprotein G [Cricetulus griseus] [MASS=42131]
99.65 0.026755 1.7 NOMO1 Nodal modulator 1 [Cricetulus griseus] [MASS=32948]
114.55 0.027062 1.5 CSE1L Exportin-2 [Cricetulus griseus] [MASS=110375]
49.38 0.027317 1.6 RANBP2 E3 SUMO-protein ligase RanBP2 [Cricetulus griseus] [MASS=342142]
40.15 0.027348 1.5 RBBP4 Histone-binding protein RBBP4 [Cricetulus griseus] [MASS=25330]
49.76 0.027401 2.2 HYPK huntingtin interacting protein K [Oryctolagus cuniculus]
65.08 0.027661 2.7 TMEM43 Transmembrane protein 43 [Cricetulus griseus] [MASS=44797]
56.29 0.027801 1.5 RRP12 RRP12-like protein [Cricetulus griseus] [MASS=153244]
48.66 0.027862 1.2 IMMT Mitochondrial inner membrane protein [Cricetulus griseus] [MASS=79258]
57.44 0.027871 1.3 EIF4B Eukaryotic translation initiation factor 4B [Cricetulus griseus] [MASS=69570]
104.08 0.02788 1.2 DSC1 rCG40478, isoform CRA_b [Rattus norvegicus]
43.46 0.028009 1.4 CTB5R3 NADH-cytochrome b5 reductase 3 [Cricetulus griseus] [MASS=34236]
47.77 0.028144 1.5 EZR Ezrin [Cricetulus griseus] [MASS=69372]
112.09 0.028961 1.2 TAGLN2 Transgelin-2 [Cricetulus griseus] [MASS=22455]
44.03 0.029486 1.4 RFC4 Replication factor C subunit 4 [Cricetulus griseus] [MASS=39938]
104.77 0.029896 1.2 EEF1G Elongation factor 1-gamma [Cricetulus griseus] [MASS=27155]
444
52.16 0.029947 1.2 VAPA Vesicle-associated membrane protein-associated protein A [Cricetulus griseus] [MASS=26075]
69.04 0.030475 1.3 PLOD2 Procollagen-lysine,2-oxoglutarate 5-dioxygenase 2 [Cricetulus griseus] [MASS=67164]
85.07 0.030576 1.5 FKBP9 FK506-binding protein 9 [Cricetulus griseus] [MASS=63008]
76.66 0.030819 1.5 RPL17 60S ribosomal protein L17 [Rattus norvegicus]
105.29 0.032149 1.3 GNB2 Guanine nucleotide-binding protein G(I)/G(S)/G(T) subunit beta-2 [Cricetulus griseus] [MASS=30483]
44.29 0.032227 1.4 PSMD3 26S proteasome non-ATPase regulatory subunit 3 [Cricetulus griseus] [MASS=16842]
57.62 0.032502 1.4 XRCC1 RecName: Full=DNA repair protein XRCC1; AltName: Full=X-ray repair cross-complementing protein 1
42.68 0.032765 1.8 RPLP1 60S acidic ribosomal protein P1 [Cricetulus griseus] [MASS=12230]
132.35 0.033095 1.3 RSL1D11 Ribosomal L1 domain-containing protein 1 [Cricetulus griseus] [MASS=49958]
43.44 0.033786 1.2 VDAC1 Voltage-dependent anion-selective channel protein 1 [Cricetulus griseus] [MASS=53379]
56.84 0.033892 1.2 LAP3 Cytosol aminopeptidase [Cricetulus griseus] [MASS=50275]
56.04 0.034082 1.4 MRPP1 Mitochondrial ribonuclease P protein 1 [Cricetulus griseus] [MASS=59932]
79.82 0.034525 1.4 VAT1 Synaptic vesicle membrane protein VAT-1-like [Cricetulus griseus] [MASS=28523]
108.57 0.034692 1.2 GTF2F1 General transcription factor IIF subunit 1 [Cricetulus griseus] [MASS=57283]
48.77 0.034812 1.7 NDUFA4 NADH dehydrogenase [ubiquinone] 1 alpha subcomplex subunit 4 [Cricetulus griseus] [MASS=14153]
61.52 0.034882 1.2 VDAC2 Voltage-dependent anion-selective channel protein 2 [Cricetulus griseus] [MASS=31739]
57.99 0.035256 1.3 GATAD2A Transcriptional repressor p66 alpha [Cricetulus griseus] [MASS=51734]
53.22 0.035477 1.2 HMGB1 RecName: Full=High mobility group protein B1; AltName: Full=High mobility group protein 1; Short=HMG-1
59.07 0.035771 1.4 DNAJC10 DnaJ-like subfamily C member 10 [Cricetulus griseus] [MASS=90672]
82.09 0.036085 1.2 IRF2BP2 Interferon regulatory factor 2-binding protein 2 [Cricetulus griseus] [MASS=36160]
46.09 0.036431 1.5 PRPF40A Pre-mRNA-processing factor 40-like A [Cricetulus griseus] [MASS=81478]
325.97 0.03688 1.9 PKM Pyruvate kinase isozymes M1/M2 [Cricetulus griseus] [MASS=51559]
62.3 0.036983 1.5 FUBP1 Far upstream element-binding protein 1 [Cricetulus griseus] [MASS=47731]
63.63 0.037078 1.2 SRSF5 serine/arginine-rich splicing factor 5 [Mus musculus]
56.67 0.037201 1.3 RCC1 Regulator of chromosome condensation [Cricetulus griseus] [MASS=48887]
48.38 0.037852 1.5 RPL31 60S ribosomal protein L31 isoform 1 [Homo sapiens]
107.52 0.037926 9.4 NCAM1 Neural cell adhesion molecule 1 [Cricetulus griseus] [MASS=114316]
42.67 0.03809 1.3 SDHB Succinate dehydrogenase [ubiquinone] iron-sulfur subunit, mitochondrial [Cricetulus griseus] [MASS=25570]
45.23 0.038162 1.4 CBR1 carbonyl reductase 1 [Cricetulus griseus] [MASS=30547]
445
42.76 0.038599 1.2 TMOD3 Tropomodulin-3 [Cricetulus griseus] [MASS=39388]
127.38 0.039203 1.2 CKAP4 Cytoskeleton-associated protein 4 [Cricetulus griseus] [MASS=36359]
140.2 0.039272 1.3 PARK7 Protein DJ-1 [Cricetulus griseus] [MASS=19929]
45.17 0.039516 1.2 HADHB Trifunctional enzyme subunit beta, mitochondrial [Cricetulus griseus] [MASS=39535]
78.61 0.039527 1.4 RINL Ras and Rab interactor-like protein [Cricetulus griseus] [MASS=71795]
59.78 0.039542 1.5 HIST1H2BC histone H2B type 1-C/E/F/G/I [Homo sapiens]
102.85 0.039719 1.2 DBN1 Drebrin [Cricetulus griseus] [MASS=65309]
43.65 0.040374 1.5 RPS23 40S ribosomal protein S23 [Homo sapiens]
172.46 0.040499 1.3 EEF1A1 elongation factor 1-alpha 1 [Rattus norvegicus]
56.47 0.041029 1.3 CLU Clusterin [Cricetulus griseus] [MASS=51757]
93.02 0.041185 2 ATIC Bifunctional purine biosynthesis protein PURH [Cricetulus griseus] [MASS=64651]
50.73 0.041398 1.9 KRT10 Keratin, type I cytoskeletal 10 [Cricetulus griseus] [MASS=27853]
83.8 0.04164 1.2 RAB5C Ras-related protein Rab-5C [Cricetulus griseus] [MASS=23483]
62.92 0.042015 2.8 RPL28 60S ribosomal protein L28 [Cricetulus griseus] [MASS=13065]
151.51 0.042153 1.2 ENO2 Gamma-enolase [Cricetulus griseus] [MASS=47227]
175.51 0.042683 1.3 CFL1 cofilin-1 [Rattus norvegicus]
88.46 0.042854 1.1 SUMO2 small ubiquitin-related modifier 2 precursor [Mus musculus]
46.05 0.043429 2.7 PRMT1 Protein arginine N-methyltransferase 1 [Cricetulus griseus] [MASS=40522]
42.13 0.043447 1.3 FDPS Farnesyl pyrophosphate synthetase [Cricetulus griseus] [MASS=43211]
269.53 0.043714 1.2 EEF1D Elongation factor 1-delta [Cricetulus griseus] [MASS=36598]
41.69 0.044283 2.1 CAV1 Caveolin-1 [Cricetulus griseus] [MASS=17160]
57.47 0.044736 1.2 RTCB UPF0027 protein C22orf28-like [Cricetulus griseus] [MASS=47794]
53.2 0.045027 1.3 CCT8 T-complex protein 1 subunit theta [Cricetulus griseus] [MASS=24801]
59.64 0.045432 1.9 AHSA1 Activator of 90 kDa heat shock protein ATPase-like 1 [Cricetulus griseus] [MASS=38129]
40.85 0.045875 1.3 IQCE IQ domain-containing protein E [Cricetulus griseus] [MASS=83194]
45.11 0.045972 1.3 FOSL1 Fos-related antigen 1 [Cricetulus griseus] [MASS=30263]
52.47 0.046094 1.3 PPM1G Protein phosphatase 1G [Cricetulus griseus] [MASS=54003]
40.31 0.046158 1.6 RARS RecName: Full=Arginyl-tRNA synthetase, cytoplasmic; AltName: Full=Arginine--tRNA ligase; Short=ArgRS
74.38 0.046454 1.2 CSPR2 Cysteine-rich protein 2 [Cricetulus griseus] [MASS=22944]
446
58.3 0.047006 1.3 PRPF31 U4/U6 small nuclear ribonucleoprotein Prp31 [Cricetulus griseus] [MASS=49130]
89.3 0.047083 1.5 ALLC putative allantoicase [Cricetulus griseus] [MASS=80219]
100.86 0.047171 3.5 TUFM Elongation factor Tu, mitochondrial [Cricetulus griseus] [MASS=49557]
53.53 0.04847 1.2 HM13 Minor histocompatibility antigen H13 [Cricetulus griseus] [MASS=37769]
49.72 0.048702 16.6 MAPRE1 Microtubule-associated protein RP/EB family member 1 [Cricetulus griseus] [MASS=30034]
41.87 0.048822 1.4 ACTR3 actin-related protein 3 [Mus musculus]
82.45 0.049156 1.1 ALDH2 Aldehyde dehydrogenase, mitochondrial [Cricetulus griseus] [MASS=45925]
80.22 0.049456 1.1 GIGYF2 PERQ amino acid-rich with GYF domain-containing protein 2 [Cricetulus griseus] [MASS=150168]
51.01 0.049711 1.2 RPS28 40S ribosomal protein S28 [Homo sapiens]
76.21 0.04984 1.3 PRDX2 RecName: Full=Peroxiredoxin-2
447
Figure A10: CT expression patterns for miR-23b/23b* in all CHO clones
Figure A10: The dogmatic miRNA expression profile of miR-23b (Red) and miR-23b* (Blue) for all
clones in the Slow growing panel compared to the Fast growing panel that exhibit an apparent
increase in miR-23b* expression with no change in miR-23b expression.
10
15
20
25
30
35
40
45
0 5 10 15 20
CT
miR-23b/23b* in Slow CHO cells
miR-23b*
miR-23b
10
15
20
25
30
35
40
45
0 5 10 15 20
CT
miR-23b/23b* in Fast CHO cells
miR-23b*
miR-23b
448
Figure A11: Raw CT values for miR-532-5p expression in all Fast and Slow CHO
clones evaluated using TLDA miRNA microarrays
Figure A11: Raw Ct values for the expression of miR-532-5p in all clones within “fast” and “slow”
growing groups (highlighted in yellow) including the average expression across each group.
Clone ID F/S Detector ID Ct Ave Ct
X5.13F.1.41 F hsa-miR-532-5p-4380928 40
X5.13F.2.28 F hsa-miR-532-5p-4380928 33.34163
X5.13F.3.31 F hsa-miR-532-5p-4380928 36.091293
X5.16F.1.40 F hsa-miR-532-5p-4380928 35.938572
X5.16F.2.28 F hsa-miR-532-5p-4380928 35.28528
X5.16F.3.33 F hsa-miR-532-5p-4380928 40
X5.18F.1.39 F hsa-miR-532-5p-4380928 32.416298
X5.18F.2.31 F hsa-miR-532-5p-4380928 40
X5.18F.3.35 F hsa-miR-532-5p-4380928 40 37.96367
X5.22F.1.44 F hsa-miR-532-5p-4380928 40
X5.22F.2.33 F hsa-miR-532-5p-4380928 40
X5.22F.3.27 F hsa-miR-532-5p-4380928 40
X5.2F.1.42 F hsa-miR-532-5p-4380928 40
X5.2F.2.31 F hsa-miR-532-5p-4380928 36.381985
X5.2F.3.25 F hsa-miR-532-5p-4380928 40
X10.2S.1.23 S hsa-miR-532-5p-4380928 22.911198
X10.2S.2.16 S hsa-miR-532-5p-4380928 24.435015
X10.2S.3.15 S hsa-miR-532-5p-4380928 23.092752
X11.1S.1.13 S hsa-miR-532-5p-4380928 32.94388
X11.1S.2.16 S hsa-miR-532-5p-4380928 40
X11.1S.3.14 S hsa-miR-532-5p-4380928 35.9087
X5.11S.1.23 S hsa-miR-532-5p-4380928 31.040468
X5.11S.2.14 S hsa-miR-532-5p-4380928 35.003376 30.99841
X5.11S.3.16 S hsa-miR-532-5p-4380928 32.213932
X5.1S.1.11 S hsa-miR-532-5p-4380928 24.3722
X5.1S.2.16 S hsa-miR-532-5p-4380928 26.736874
X5.1S.3.19 S hsa-miR-532-5p-4380928 24.942024
X7.7S.1.17 S hsa-miR-532-5p-4380928 40
X7.7S.2.20 S hsa-miR-532-5p-4380928 36.03121
X7.7S.3.15 S hsa-miR-532-5p-4380928 35.344524
449
Figure A12: miR-18a expression in a stable miR-17~92 cluster over-expressing
CHO-K1 parental cell line
Figure A12: The expression of miR-18a was evaluated in a parental CHO-K1 parental cell line stably
expressing the miR-17~92 cluster.
Figure A13: endogenous miR-23b* levels in miR-23b* sponge CHO-SEAP cells
Figure A13: qPCR assessment of the expression of endogenous miR-23b* in CHO-K1-SEAP clones
engineered to stably deplete miR-23b*.
0
0.5
1
1.5
2
2.5
CEH empty CEH-miR-17-92
RQ
miR-18a
miR-18a
**
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
miR-NC-2 miR-NC-11 miR-NC-20 miR-23b*-18
RQ
miR-23b*
450
Figure A14: CHO-SEAP miR-23 clones 4 and 6 transfected with pre-miR-23
Figure A14: CHO-SEAP clones 4 and 6 stably depleted of miR-23 transfected with pre-miR-23 mimic
demonstrating A) the GFP readings from the real-life Guava ExpressPlus programme and B) the
average GFP expression within each population.
pre-miR-NC pre-miR-23b
0
200
400
600
800
1000
1200
1400
pre-miR-NC pre-miR-23b pre-miR-NC pre-miR-23b
GF
P E
xp
ress
ion
miR-23b spg 4/6 GFP exp
GFP
miR-23b spg 4 miR-23b spg 6
***
***
A
B